Skip to main content

Full text of "Roasting of gold and silver ores : and the extraction of their respective metals without quicksilver"

See other formats


311 073 






THE ROBERT E. COWAN COLl. 

I'RKSKNTICI) TO Till. 

UNIVERSITY OF CALIFORNIA 



C. P. HUNTINGTON 

dUNE. 1897. 

Recession H.Q...^ Class No, 







ROASTING 



OF 



GOLD AND SILVER ORES 



AND THE 



EXTRACTION 



OF THEIR 



RESPECTIVE METALS 

WITHOUT QUICKSILVER, 



BY Cr. liUSTEL, 

MINING ENGINEER AND METALLURGIST, 

Author of " Nevada and California Processes of Silver and Gold Extraction, 
and " Concentration of all Kinds of Ores." 



Illustrated with Numerous Engravings. 




PUBLISHED AND SOLD BY 

DEWEY & CO., PROPRIETORS SCIENTIFIC PRESS. 

SAN FRANCISCO, 1870. 



Entered according to Act of Congress, in the year 1870, 

BY G. KUSTEL, 

In the Clerk's Office of the District Court of the United States for 
the Northern District of California. 



Printed by SPAULDING & BARTO, 

SCIENTIFIC PRESS BOOK AND JOB OFFICE, 

414 Clay Street, San Francisco. 




PREFACE. 



The publication of this Treatise is due solely to the many 
inquiries concerning the "Leaching, Solving and Precipita- 
tion Process for Silver Ores, ' ' now successfully practiced in 
Sonora, Mexico, where it has been lately introduced by Mr. 
Ottocar Hofmann. 

In consideration of the very important preparation of the 
ore, before it is subjected to the Solving Process, namely, 
the Roasting, I have thought it proper to devote consider- 
able space to the description of different modifications of 
this operation, which is regulated by the peculiarity of the 
ore, and by the subsequent treatment. It is impossible to 
give any one way which will be suitable in every case ; for 
this reason, and in order to cover all cases as far as possible, 
a detailed description of different modes of Boasting will not 
be superfluous. 

The Solving Process, as now practiced, is a very econom- 
ical method for the extraction of silver, for the reason that no 
quicksilver and no castings are used except what are needed 
for crushing. Mills in Mexico being dependent on San Fran- 
cisco for the shoes, dies, gearing, etc., of amalgamating pans, 
millmen there know how to appreciate a process confined to 
wooden tubs requiring no power. A comparatively small 
capital is necessary for building up such works, and hence 
there is a more reasonable ratio established between the 
amount of money which must be expended on the works and 
the real value of the mine, than where other more expensive 



IV PREFACE. 

machinery is employed, a circumstance which, being insuf- 
ficiently regarded, is often the source of failure. 

Mr. O. Hofmann commenced first with the " Chlorination 
Process" (82), but finding great difficulty in obtaining the 
regular supply of sulphuric acid and manganese from San 
Francisco, abandoned the chlorination with cold chlorine gas, 
which is indispensable in the presence of gold. Another dif- 
ficulty was in obtaining a good article of sulphide of sodium. 
He tried to extract the potash from ashes, and to use this in 
place of soda, but decided finally in favor of lime, which is 
found in abundance. From this the sulphide of calcium is 
easily manufactured on the spot. Sulphide of calcium was 
first applied by Kiss ($80). 

The Solving Process is very simple, and readily performed 
by common w r orkmen ; besides the lime, only brimstone 
must be provided, in order to prepare the necessary chem- 
icals for solving and precipitation. It is a general but 
erroneous belief, that the solving is a slow process. An 
amalgamating pan is charged with 500 to 1,000 pounds of 
roasted ore, and treated at least six hours, and therefore 
turns out at most two tons in 24 hours ; while a box or vat 
of proper size used in the Solving Process, can work from 
four to five tons in the same time. 

Only those ores are treated by this process which abso- 
lutely require roasting ; which, however, with improved fur- 
naces, is not so expensive as it used to be. The chloride 
ores alone can be leached directly without roasting, and this 
when there is no other silver combination in them. 

G. KUSTEL. 

MABCH, 1870. 




I. INTRODUCTION. 



Classification of Ores. 

1. Ores ma} 7 be classified : . According to 
the metal, the extraction of which is principally 
remunerative ; as silver ores, lead ores, copper 
ores, etc. b. According to the metallurgical treat- 
ment ; as roasting ores, smelting ores, amalgamat- 
ing ores, etc. c. According to the predominant 
gangue, as calcareous ores, quartzose or ochery 
ores. d. According to the predominant metallic 
mineral; as sulphuret ores, chloride ores, carbon- 
ates, etc. 

Important Silver Ores. 

2. The most important silver ores are those 
found in such quantities as to be an object of 
metallurgical operations. The principal minerals 
of this kind are the following : 

A. Real Silver 0?*es. a. Sulphuret of Silver, or 
silver glance, with 87 per cent, of silver. It is of 
common occurrence, and is the most suitable of the 
silver sulphurets for pan amalgamation without 



6 INTRODUCTION. 

roasting, b. Brittle Silver Ore, or sulphuret of 
silver and antimony. This mineral contains 68 per 
cent, of silver, and is quite common, c. Polyba- 
site, sulphuret of silver, antimony and some arsenic, 
with 75 per cent, of silver. Brittle silver ore and 
polybasite are both tractable in pans without roast- 
ing, although not so readily as the simple sulphuret. 
All other sulphureted silver ores require roasting. 
d. Ruby Silver. The dark red silver ore, or antimo- 
nial variety, with 59 per cent., and the light red 
silver ore, or arsenical variety, with 65 per cent, 
of silver, are valuable minerals. They occur quite 
frequently in Nevada, Idaho, Montana, Mexico, 
etc. e. Miargyrite, sulphuret of silver and anti- 
mony; 36.5 per cent, of silver; Idaho, Montana, 
etc. /. Stromeyerite, or silver copper glance, a 
sulphuret of silver and copper containing up to 53 
per cent, of silver; Nevada, Arizona, etc. g. 
Horn Silver, or chloride of silver, with 75 per cent, 
of silver; occurs massive in White Pine, Nevada; 
prepared by nature for the pan amalgamation, h. 
Stetefeldtite and Partzite, with up to 25 per cent, of 
silver, are oxide ores which occur very frequently 
in Nevada, Arizona, etc. 

B. Argentiferous Ores. a. Silver-fahl-ore , ar- 
gentiferous gray copper ore. It contains silver in 
very variable proportions up to 31 per cent. This 
ore is quite common, and for this reason is im- 
portant. It is also one of the most rebellious ores, 
containing copper, antimony, arsenic, sulphur, 



INTRODUCTION. I 

lead, iron, zinc, and sometimes gold and quick- 
silver, b. Argentiferous Lead Ores, galena, or sul- 
phuret of lead, lead glance. Generally, this is not 
rich in silver, containing from $20 to $60 per ton. 
Specimens assay sometimes as high as $300. The 
fine grained variety is generally considered richer 
than the coarse crystallized kind, but this has not 
been observed to be the case in Nevada and Ari- 
zona, c. Cerusite, carbonate of lead. If pure, 
without admixture of copper and other carbonates, 
it is poor in silver in most cases. Raw, it amalga- 
mates only too readily in pans. Smelting is the 
only proper way of treating galena and cerusite. 
d. Argentiferous Zincblende. Sulphuret of zinc. 
Pure zincblende contains usually only traces of 
silver; often, however, it assays well, even up to 
$400 per ton, although no other silver ore can be 
detected with it. In some mines the argentiferous 
zincblende prevails, and is the most important ore. 
It requires a great heat in roasting, e. Argentifer- 
ous Pyrites. Copper and iron pyrites are poor in 
silver, but often auriferous. Pyrite is a valuable 
companion for silver ores which have to be treated 
by a chloridizing roasting, on account of its amount 
of sulphur, which is necessary for the decom- 
position of salt. 

Difference between Real Silver Ores and 
Argentiferous Ores. 

3. Heal silver ores have mostly an unvariable 
amount of silver. Heal silver minerals admit an 



8 INTRODUCTION. 

approximate estimate of the value of the ore, if the 
proportion of ore and gangue is considered, with- 
out making an assay. With the argentiferous ores 
it is different. Fahl ore, for instance, may be very 
poor or very rich, 'and its value can be ascertained 
only by an assay. There are no means of esti- 
mating the richness of argentiferous ores "by 
sight." 

Important Combinations. 

4. With the exception of a few metal oxides of 
iron, zinc, tin, manganese, and, among silver ores, 
of the stetefeldtite, etc., the most important, be- 
cause most frequent ores, are the sulphureted vari- 
eties. Sulphur is the most formidable obstacle to 
the metallurgist in extracting metals from their re- 
spective ores. Desulphurization has been a sub- 
ject of most diligent and numerous experiments. 
The oldest method is the application of heat, which 
is still in use, notwithstanding the many attempts 
in modern times to dispense entirely with fire or to 
modify its application so as to perform the process 
more perfectly and in a shorter time. 

Means of Desulphurization. 

6. The desulphurization of ores is effected: a. 
By heating with free admission of air. This is the 
common way of " roasting," and the most impor- 
tant, and is effected either in kilns, heaps, etc., or 



INTRODUCTION. U 

in reverberatory furnaces. As soon as the sulphu- 
rated ore is heated to a certain degree, one part of 
the sulphur escapes as sulphurous acid; another is 
converted into sulphuric acid, which is also decom- 
posed by an increased heat. Some sulphurets 
(iron pyrites) lose their sulphur without the appli- 
cation of heat, being decomposed by exposure to 
the action of air for a long time. This way is 
sometimes practiced on gold-bearing pyrites, b. 
By heating with exclusion of air. Only the sul- 
phides of gold and platinum are decomposed per- 
fectly by this method. Other sulphureted ores 
lose their sulphur only in part, being reduced to a 
lower state of sulphide. Sulphuret of silver (Ag 
S) remains undecomposed. Cinnabar, sulphide of 
antimony (Sb S 3 ) and sulphide of arsenic volatilize 
unchanged. Iron pyrites (Fe S' 2 ) gives up 23 per 
cent, of its sulphur, being reduced to magnetic 
pyrites, and, by a strong heat, to proto-sulphide of 
iron (Fe S), not further reducible. Also sulphide 
of zinc (zincblende), remains undecomposed. 
Copper glance retains its sulphur, and copper py- 
rites loses only one part of the sulphur which 
is combined with the iron in it. Galena (Pb S) is 
reduced to a lower state (Pb 4 S), a part of the lead 
separating out in a metallic state, c. By super- 
heated steam. Sulphurets not evolving sulphur by 
the last process, lose their sulphur slowly on the 
application of steam, sulphureted hydrogen and 
sulj hurous acid being formed. Experiments made 
by Begnault showed that desulphurization is 
1* 




10 INTRODUCTION. 

effected more perfectly if air is admitted. Roast- 
ing in reverberatory furnaces is always effected "by 
the oxygen of the air and by steam, as there is no 
fuel used which contains less than 25 to 30 per 
cent, of water. Superheated steam has been tried 
in different ways on sulphurets with the highest 
expectations, but with no better results for practical 
use than are given in the ordinary way by the 
steam obtained from fuel. It may be useful in 
many instances to have more steam than is thus 
obtained, but this increases considerably the ex- 
pense of roasting; as, for instance, in Patera's ap- 
plication of steam in roasting silver ores, tried 
principally with the intention of expelling anti- 
mony, arsenic, etc. Another application of super- 
heated steam, with exclusion of air, is Hagan's 
method, which may prove successful on pyritous 
ores, having at the same time the advantage of 
being a very cheap method, d. By heating with 
metals, alkalies or alkaline earths, for which the 
sulphur has a greater affinity. The affinity of sul- 
phur for the following metals decreases in the 
order in which they stand, being strongest for the 
first and weakest for the last: Copper, iron, tin, 
zinc, lead, silver, antimony, arsenic. Each of 
these metals can be desulphurized by the next pre- 
ceding, though with difficulty; but more easily by 
one further off. Practical use of this property is 
made in smelting galena with the addition of me- 
tallic iron or iron ore. Sulphide of silver in cru- 
cibles is decomposed by stirring the liquid with 



INTRODUCTION. 11 

red hot iron. Quicksilver is obtained from cinna- 
bar by heating the latter with lime, which takes up 
the sulphur, etc. e. Carbon has no great affinity 
for sulphur; the use of charcoal for desulphurization 
of ores is therefore an inferior method. So is also 
the use of carbonic acid. 

Result of Desulphurization. 

6. The direct extraction of metals from sul- 
phurets, either by smelting or amalgamation, is 
not practicable. In smelting, the sulphurets melt 
very readily, but only a small part, if any, of the 
metal is obtained, while the greater part runs out 
combined with the sulphur as matt. For this reason 
the roasting of sulphureted ore for the purpose of 
smelting is indispensable, unless iron is added. 
Such roasting or burning takes often many weeks, 
or months. The direct amalgamation, also, of sul- 
phurets gives a veiy poor result, except in the case 
of silver glance. By means of electricity, com- 
bined with the chemical action of sulphate of cop- 
per and salt, the silver and gold sulphurets are de- 
composed; but, with the exception of the patio 
amalgamation, no process has yet been publicly 
demonstrated as really practical for the treatment 
of all kinds of raw sulphurets. The desulphuriza- 
tion is therefore still a most important preparation 
for the extraction of metals. The general effect of 
roasting is that the metals are oxidized. Only 
gold and silver are transformed into a metallic 



1 2 INTRODUCTION . 

condition; and of the silver, moreover, a large per- 
centage is always found as a sulphate, even when 
the roasting is well performed. Some of the silver 
combines as an oxide with antimony and silica, if 
present. All the oxides obtained by desulphuriza- 
tion must be again deoxidized in order to get them 
in a metallic state. 

Means of Reduction or Deoxidation. 

7. Heating alone will reduce the oxides of the 
precious metals only. Oxide of gold does not 
occur in nature, neither is it obtained in any of the 
metallurgical processes. Oxide of silver is also 
unimportant; it is formed, to a small extent, in cu- 
pellation (and taken up by the litharge), in smelt- 
ing silver ores combined with silica, and in roasting 
silver ores in the presence of antimony, arsenic, 
etc. 

The most powerful agent of reduction is carbon 
(charcoal, coke, etc.) and carbonic oxide. In all 
smelting in blast furnaces, the carbonic oxide is 
the real reducer. The burning coal, under the in- 
fluence of the compressed air, produces carbonic 
acid, melting at the same time the ore; the car- 
bonic acid, passing through the glowing coal above 
the melting region, gives up a part of its oxygen to 
the coal, and is reduced thereby to carbonic oxide, 
which in turn takes up oxygen again, from the 
metal oxides, reducing them to a metallic state; a 
contact of ore oxides with carbon is therefore not 



INTRODUCTION. 13 

necessary for the purpose of reduction. All metals 
do not retain their oxygen with equal tenacity, but 
some part with it much more easily than others. 
For instance, lead, copper, bismuth, antimony, 
cobalt and nickel, require for their reduction a 
darker or lighter red heat, while iron, zinc and tin 
are reduced only at a white heat. But also hydro- 
gen and carbureted hydrogen, created by the 
burning fuel, are powerful reducing agents. 

Metal oxides in solution are reduced and precip- 
itated in a metallic condition by other metals. On 
this principle copper is precipitated by metallic 
iron, which goes into solution in place of the cop- 
per; sulphate of silver, in Ziervogel's process, is 
precipitated by copper, etc. Also, by aid of the 
electro-galvanic stream, metals are reduced to a 
metallic state from their solutions. 

Desulphurization of Silver Ores not Effi- 
cient. 

8. Although, by mere desulphurization the 
silver is to a great extent converted into a metallic 
state, this is not always its most suitable condition 
except for smelting. Almost all silver extracted in 
the United States is obtained by amalgamation, 
smelting being confined to a few localities where 
the ore contains such a high percentage of lead 
that its amalgamation is impossible. It would 
seem as if metallic silver should amalgamate more 
easily than if combined with another substance. 



14 INTRODUCTION. 

t 

This, however, is not the case. The silver, after 
roasting, is generally coated with the oxides of 
volatile base metals, which prevent its ready amal- 
gamation. Moreover, a direct contact between 
quicksilver and silver is a necessary condition for 
their amalgamation. A momentary contact in a 
muddy pulp is not always successful. The chlo- 
ride of silver, however, goes into solution and 
unites easily with the quicksilver. Hence, in most 
instances, it is necessary to adopt a chloridizing 
roasting. 

What a Chloride is, and how Chlorination 
is Effected. 

9. The term chloride is applied to all com- 
pounds of chlorine with a metal or other radical. 
Chlorine is a greenish-yellow gas, an elementary 
substance, of 2.45 specific gravity, and of a peculiar 
and disagreeable odor. It is not found free in na- 
ture, but always in combination, principally with 
sodium, forming common salt. Metallic chlorides 
are of frequent occurrence. Chlorine is, for in- 
stance, combined with silver as horn silver, with 
copper as Atacamite, with lead as Kerasine, Mendi- 
pite, etc; also with quicksilver as Calomel. 

10. To chloridize ore, that is, to convert the 
metals into chlorides, it is necessary to produce 
chlorine and to bring it in intimate contact with 
the ore particles. The cheapest material evolving 



INTRODUCTION. 15 

chlorine is salt (chloride of sodium), and the only 
practical way of separating the chlorine from sodi- 
um is by substituting for it another substance for 
which the sodium has a stronger affinity. The 
cheapest ingredient for this purpose is sulphuric 
acid. The sodium being oxidized to soda, unites 
with the sulphuric acid, forming sulphate of soda, 
while chlorine is set free. 

For the treatment of ores there are two principal 
methods of chloridizing. One is roasting the ore 
with salt in a furnace; the other is the " cold chlo- 
rination." Boasting, at first, when in the presence 
of salt, has an oxidizing effect, as there is then no 
sulphuric acid present to decompose the salt, and 
the heat alone would, if increased, volatilize and 
not decompose this. The sulphurets in the ore, 
under the influence of heat, lose a part of their 
sulphur as sulphurous acid gas; the other part of 
the sulphur oxidizes to sulphuric acid. As soon as 
this is formed it attacks the salt, and the chlo- 
rine, being set free, then acts on metals, metal 
oxides, sulphurets, arseniurets and antimgnial com- 
binations, forming partly metal chlorides and 
partly chlorides of sulphur, arsenic and antimony. 

11. The other mode of chloridizing consists in 
the employment of cold chlorine gas with roasted 
ores, principally desulphurized gold ores, but of 
late,. also silver ores. The chlorine must be pro- 
duced here separately, and conducted into the cold 
ore by leaden or india rubber pipes. The ingredi- 
ents are : salt, sulphuric acid and peroxide of man- 



16 INTRODUCTION. 

ganese. Salt is first attacked by the sulphuric 
acid, and hydrochloric acid and sulphate of soda 
are formed. The hydrogen of the hydrochloric 
acid then combines with the oxygen of the manga- 
nese, and the chlorine escapes. A part of the 
chlorine unites with the manganese, but is decom- 
posed again by sulphuric acid, so that all chlorine 
is expelled from the salt, leaving sulphates of soda 
and manganese in the gas generator. The chlori- 
nation of gold, unlike that of silver, is difficult to 
effect in a furnace ( 38), for the reason that, if 
formed, the gold chloride is reduced back to the 
metallic state at a low, almost dark red heat. The 
difference between hot and cold chlorination is 
principally found in the fact that, while in the first 
way a great many base metal chlorides are formed, 
the cold chlorine combines principally with the 
free metal, with silver and gold; while the other 
metals, being oxidized, are not decomposed by the 
chlorine. Silver oxide, if present, is decomposed 
and chloridized. 

Chlorination is also effected by chemical decom- 
position in the wet way, as practiced in the Mexi- 
can patio amalgamation, by mixing with the ore 
sulphate of copper and salt. 

Means of Separating the Metal from Chlo- 
rine. 

13. The chloride of silver can be melted with- 
out being altered; chlorides of gold and of plati- 



INTRODUCTION . 1 7 

num. lose all their chlorine on being heated. 
Chloride of iron exposed to air and heat, as is the 
case in a chloridizing roasting, loses its chlorine 
and is changed to iron oxide. The chloride "of 
copper gives up only a part of its chlorine. Heat- 
ing alone has therefore no practical value for the 
disengagement of chlorine. 

The most effective way of separating the chlorine 
from the metal is the application of another metal 
for which the chlorine has more affinity. On this 
property of chlorine is based the amalgamation of 
silver ores, after a chloridizing roasting, in pans, 
tubs and barrels, and the patio amalgamation. 
The chloride of silver in the ore is decomposed, 
and the silver set free during amalgamation in iron 
pans by the metallic iron of the pan, or if quick- 
silver is charged at the same time with the ore, by 
both the quicksilver and the iron. In the barrel 
amalgamation the silver is disengaged by metallic 
iron, and in the patio amalgamation by quicksilver. 
In all these instances the silver, being deprived of 
its chlorine, alloys with the quicksilver and forms 
the amalgam. 

On the same principle the metal is extracted 
from soluble chlorides. The proto-chlorides are 
all more or less soluble in water, except that of sil- 
ver, which is quite insoluble. The chloride of 
copper, in solution, is brought together with me- 
tallic iron, or conveyed over it. The chlorine of 
the copper unites with the iron, and the copper 
falls in a metallic state, ready to be melted into 



18 INTKODUCTION. 

bars, after being washed, pressed and dried. In- 
directly, the silver is extracted from its chloridized 
state by dissolving the chloride in the hyposul- 
phites of soda, of potash, or of lime. In these salts 
the silver chloride dissolves very readily, giving a 
clear solution of a very sweet taste, out of which 
the silver is precipitated by the corresponding 
alkaline sulphides as sulphide of silver. 

The chloride of gold, obtained from the chlo- 
rination of gold-bearing sulphurets, is precipitated 
by sulphate of iron in such a way that metallic gold 
results, while the chlorine combines with a part of 
the iron. 

13. The silver is easily obtained from the chlo- 
ride by melting it with alkalies; for instance, with 
soda, potash or lime. The chlorine unites with 
sodium, calcium, etc., and the silver separates on 
the bottom of the crucible. If there is not a suffi- 
cient amount of the alkalies present, some silver 
will be lost. In most instances it is preferable to 
mix the artificial chloride with water and some sul- 
phuric acid and granulated zinc, or] zinc sheet if 
smaller quantities are being operated on. The 
chloride of silver by degrees changes its white., 
color to a dark gray, being converted into the me- 
tallic state in a short time. It is reduced to 
metal by the nascent hydrogen. After the sulphate 
of zinc, which is formed and dissolved, has been 
washed away, the silver is pressed, dried, and, with 
addition of some soda and borax, melted into a 
bar. 



'ROASTING OF ORES. 19 

In the same way as from a sulphate, silver can 
be precipitated by copper, after the chloride of 
silver has been dissolved in a hot solution of salt, 
as is done in Augustin's process. This is not prac- 
ticable with the argentiferous solution of hyposul- 
phite of soda. 

By using sodium amalgam and iron filings, the 
silver chloride is instantly decomposed and silver 
amalgam formed. 

The chloride of gold is precipitated in a metallic 
condition; also by the chloride of iron (Fe Cl), the 
consideration of which is important in treating sul- 
phurets by chlorination. 



II. ROASTING OF ORES. 



14. The object of roasting is either to effect 
chemical changes, as required for amalgamation, 
smelting, etc., or sometimes also to reduce the 
hardness of the ore, in order to make it easier to 
crush. Roasting for the latter purpose, exposing 
the ore to the fire in large pieces, is more properly 
termed " burning." The beginning of smelting is 
under all circumstances beyond the limits of roast- 
ing; therefore all roasting furnaces in which the 
regulation of heat is so far out of the control of the 



20 BOASTING OF ORES. * 

roaster that a partial smelting would arise, are un- 
fit for roasting. This is often the case with vertical 
furnaces. But although a partial smelting or clot- 
ting is not within the province of roasting, and in 
all instances is very injurious to the result of sub- 
sequent amalgamation or precipitation, it is never- 
theless applied with much success on concentrated 
ore intended for smelting. By this process the 
loose sand assumes a compact form, the gases and 
wind penetrate the charge more easily, and the loss 
in metal is diminished. 

If there is no necessity for effecting a perfect 
chemical change in the ore, or if roasting is re- 
quired for smelting purposes, and a powdered form 
is not admissible, the ore is taken in larger or 
smaller pieces, generally not below the size of a 
hen's egg, and subjected to roasting either in open 
heaps, in kilns or in vertical or reverberatory fur- 
naces. In roasting in heaps, the wood is first 
placed on the ground, sometimes surrounded by a 
wall two or three feet high, then the ore is put 
over it. Less frequently ore and wood are laid 
in strata. If there is sufficient sulphur in the ore, 
the burning will continue without addition of fuel 
for many days or weeks. It is evident that the re- 
sult of such roasting is very unequal, the outside 
being more oxidized than the inside, the heat 
greater near the fuel than further off, etc. For 
this reason such ore is often roasted over several 
times. 

In vertical furnaces, the ore is laid in strata alter- 
nating with fuel, or there are several fire-places 



ROASTING OF ORES. 21 

outside the furnace so arranged that the flame 
is conducted by the draft into the furnace. A 
modification in construction and principle is the 
Hagaii roasting furnace, in which the decom- 
position of superheated steam is a source of cre- 
ating heat and a decomposing agent at the same 
time. The roasting is performed in a short time, 
and with proper ore and pieces of the right size 
the result is very satisfactory. It is also a cheap 
process/ and is applied for roasting gold-quartz 
holding sulphurets, the amalgamation of which, 
without roasting, is defective. This kind of roast- 
ing would be also applicable as preparatory for 
amalgamating silver ores or for the chlorination 
process ( 74). 

In most instances with silver ores, a most perfect 
chemical change is a condition on which the result 
of extracting the silver depends ; and for this pur- 
pose the ore must be pulverized, in order to effect 
a perfect contact between ore particles, gases, and 
other substances which are mixed with the ore for 
certain purposes. The roasting of the pulverized 
ore is executed mostly in reverberatory furnaces; 
sometimes, also, in a kind of retort furnace, if the 
roasting should be done without the admission of 
air. There are also other furnaces lately intro- 
duced or tried, the description of which will be 
found hereafter. 

In accordance with the intended mode of ex- 
traction, the ore is either roasted with an addition 
of charcoal powder, whereby the silver is reduced 



22 BOASTING OF ORES. 

to a metallic state, a procedure of little practical 
use, or the ore is subjected to an oxidizing roast- 
ing, with the principal object of driving out 
arsenic, antimony or sulphur, converting at the 
same time the silver into a sulphate (Ziervogel's 
process); or a chloridizing roasting is effected, that 
is, roasting with salt. 

A. Chloridizing Roasting. 

^ 

15. In order to chloridize the ore, an addition 
of common salt is indispensable. The salt furnishes 
chlorine for that purpose, and is decomposed by 
sulphuric acid. The sulphuric acid is created by 
the decomposition of sulphurets present in the 
ore (10). .It follows that if silver ore is to be 
roasted successfully with salt, there must be a cer- 
tain percentage of sulphurets in it; otherwise no 
sulphuric acid can be obtained, and consequently 
no chlorination, or at least only a very imperfect 
one, can be effected ( 18). 

Before introducing the ore into the furnace the 
latter must be gradually heated up, which may take 
ten to fifteen hours. When nearly red hot, a charge 
of dry ore, mixed with salt, is brought on the hearth 
through the roof and spread out equally by means 
of a hoe. The fire is kept up moderately, but suf- 
ficient flame must be seen over the ore. The draft 
is lessened by the damper, and the ore stirred dili- 
gently, but not continually. The intervals, how- 
ever, must be short. In case the ore contains lead 



OP TH3 

UNIVERSITY 



BOASTING OF ORES. 



and antimony, it is advisable to stir continually 
for at least three hours. The ore by degrees be- 
comes red hot, and the burning of the sulphur is 
quite lively. One part of the sulphur, by the 
action of oxygen, is converted into sulphuric acid 
and combines with the metals, deprived of their 
sulphur or arsenic, to a sulphate. The period of 
the formation of sulphates is very important and 
requires some time before it is finished. If there 
is a large amount of sulphurets in the ore, the 
burning of the sulphur creates so much heat that 
the feeding of the fire must be stopped almost en- 
tirely for an hour or two, but must be resumed 
again as soon as it is perceived that the ore com- 
mences to cool. The workman stirs the ore, with 
a hoe or an iron rake, back and forward across 
the hearth, moving it from the bridge toward the 
flue and back. The formation of sulphates still 
continues with disengagement of sulphurous gas. 
The ore at the bridge is more exposed to heat than 
that on the opposite side, and the roaster is obliged 
to change the ore by raking it together into a long- 
heap extending from the bridge toward the flue 
not in the middle, but nearer the working door. 
By means of a shovel, six inches by twelve, on a 
long (12-foot) iron handle, the roaster takes the ore 
from near the bridge and transfers it toward the 
flue, putting it behind the ridge of ore until he 
reaches the middle of the furnace. He then takes 
the other end of the ridge and moves it toward 
the bridge. After this the stirring is continued in 
the usual way. 



24 



ROASTING OF ORES. 



16. The sulphates react now on the salt, and 
decompose it under increased heat, setting the 
chlorine free. A mutual exchange takes place in 
part. Sulphate of lead changes into chloride of 
lead, which, volatilizing and coming in contact with 
air, loses one part of its chlorine and is reduced to 
a combination of oxy-chloride of lead. Sulphate 
of iron and sulphate of copper change also into 
chlorides. The copper chloride becomes volatile, 
colors the flame blue, emits chlorine gas and forms 
subchloride of copper. The chlorine, set free, de- 
composes the sulphurets and sulphates of silver, 
and creates chloride of silver. If, during the op- 
eration, lumps are formed, in case the ore was not 
dry enough or too much heat was applied in the 
beginning, they must be crushed to powder by 
a hammer-like iron instrument with a long handle. 
As soon as the chlorination begins, after three or 
four hours, a different smell, that of chlorine, will 
be observed. White fumes arise, and gases and 
vapors are evolved, consisting of sulphurous acid, 
chlorine, hydrochloric acid gas, chloride of sulphur, 
of iron and of copper. 

The ore increases now in volume and assumes a 
wooly condition. Another hour's roasting will 
now finish the chlorination. This last hour's stir- 
ring requires a light red heat in order to destroy as 
much as possible of the base metal chlorides. If 
there is a great percentage of copper and other 
base metals in the ore, the roasting may require 
more time, in order to decompose the chlorides 



ROASTING OF ORES. 



25 



and sulphates, the presence of which consumes too 

much iron, and during amalgamation in barrels, 
increases the heat to such a de- 
gree as to cause an injurious di- 
vision of the mercury into small 
particles and scum. The base 
metal chlorides are reduced by 
the iron and also amalgamated. 

The changing of the cooler por- 
tion near the flue with the hotter 
part at the bridge must be re- 
peated two or three times during 
the roasting process. When fin- 
ished, after five or six hours, the 
ore is drawn out and discharged 
through the discharge-hole in the 
bottom. White fumes and gases 
are still arising. 

The hoe, Fig. 1, is made of 
J-inch wrought iron, six inches 
high and eight inches wide. The 
rod or handle must be fifteen feet 
long at least. This would render 
the instrument heavy and tiresome 
to handle ; it is therefore prefera- 
FIG. 2. ble to use a piece of gas-pipe, 
welding it together with the rod 

as represented in Fig. 1. The rake is generally 

of cast iron, and is shown in Fig. 2. 
2 




26 ROASTING OF ORES. 



Necessary Amount of Sulphurets. 

17. In times when the barrel amalgamation 
was yet practiced in Freiberg (Saxony), long expe- 
rience showed that a large amount of iron sulphu- 
rets was necessary in order to decompose the 
amount of salt required for the chlorination. One 
hundred parts of the ore were mixed with 150 parts 
of borax glass, 100 parts of common glass and one 
part of resin. This mixture, melted in an assay 
crucible, gave a button of matt (sulphide of iron), 
the weight of which was from 25 to 30 per ent. of 
the original weight of the ore. If less matt was 
obtained, the ore was considered too poor in sul- 
phurets and more pyrites had to be added. 

There is not much silver ore found in the State 
of Nevada which would give 25 per cent, of matt 
on the average; and as there is no pyrites to be 
obtained for this purpose, the ore must be roasted 
as it is. When starting the first amalgamation 
works in Nevada, I found from six to eight per 
cent, of sulphurets (different kinds) in the Corn- 
stock ore, which, after roasting, contained 88 per 
cent, of its silver converted into a chloride. The 
ore from the Rising Star mine (Idaho) had not over 
8 or 10 per cent, of sulphurets, still there was 91 
per cent, of chloride of silver found after roasting. 
It is, however, very probable that from silver ores 
containing a great deal of calc spar or heavy spar, 
a less satisfactory result might be obtained by 



ROASTING OF ORES. 27 

chloridizing roasting, if no more than 6 per cent, of 
sulphurets should occur in them. Some copper- 
holding ores, especially if other base metals are 
present, and no sulphur (or very little), will give 
sometimes a good chloridizing roasting without 
any addition of green vitriol or other sulphur com- 
bination. 

18. In treating ores entirely free from, or with 
a very small percentage of sulphurets, the want of 
sulphuric acid must be remedied by adding another 
substance. A cheap material of this kind is found 
in the green vitriol or copperas (sulphate of iron), 
of which lij to 3 per cent, is added when 8 to 10 
per cent, of salt is used. The copperas is first 
calcined, in order to drive out its water fcf crys- 
tallization, by a gentle heat, and from the calcined 
article, not the crystallized, is taken the above 
percentage. This sulphate acts then on the salt 
the same as if it were created in roasting. The 
copperas is also added to arsenical ores free from 
sulphurets. But the percentage of green vitriol 
to be added depends also on the nature of the 
gangue. If there is a great deal of lime in the ore 
it takes up sulphuric acid, forming sulphate of 
lime, remaining in this condition through the pro- 
cess of roasting without being decomposed further. 
For this reason calcareous ore requires as much 
more green vitriol or iron pyrites as is necessary to 
transform all lime into a sulphate. Silica or 
quartz, if abundant, in the presence of steam, de- 



28 ROASTING OF ORES. 

composes some of the salt when red hot, forming 
silicate of soda and hydrochloric acid, the impor- 
tance of which is shown by the fact that gaseous 
hydrochloric acid, in contact with metallic silver, 
unites with it to a chloride. It behaves in a like 
manner with sulphur ets and arsenides, of which the 
most are decomposed, forming chlorides, while 
sulphur and arsenic escape combined with hy- 
drogen. 

Amount of Salt to be Used, and When. 

19. Ores containing from 80 to 100 ounces of 
silver per ton should be mixed with 10 per cent, of 
salt. This is about the quantity considered neces- 
sary iiPthe amalgamation works of Europe. Rich 
ore is often roasted with 20 per cent, of salt. If 
all the chlorine of the salt could be transferred to 
the silver, an insignificant amount of salt only 
would be required for ores containing 100 ounces 
of silver not more than 3J pounds to the ton; but 
in consequence of the different ways in which 
the chlorine decomposes and unites with base 
metals and gases, the escape of chlorine from the 
surface of the ore without coming in contact with 
the silver, etc., a great deal more of the salt must 
be applied. 

The usual amount of salt used in the United 
States for ores of the above value, is from 120 to 
140 pounds per ton of ore; that is, from 6 to 7 per 
cent. It is not advisable to take less than 6 per 



ROASTING OF ORES. 29 

cent. ( 37), even if the ore be poorer. There are 
instances, however, where 1)1 per cent, of silver 
has been obtained by amalgamation from ores 
which were roasted with only 5 per cent, of salt. 
There was no natural chloride of silver in the ore 
when treated with 5 per cent. (Rising Star ore). 

As the salt is not at all decomposed before the 
formation of sulphates commences, or only to a 
very small extent, it is also in this respect immate- 
rial whether the salt is charged at once with the 
ore, or whether it is introduced two hours later, 
unless the ore is of such a nature as would bake 
easily on a little increase of heat. In other 
cases, however, it is obvious that, taking only 6 
per cent, of salt, and employing only one man at a 
furnace, a perfect mixing in a short time, as ought 
to be done if the salt is charged after the sulphur 
is burned off, cannot be expected, and conse- 
quently a defective result will follow. It is there- 
fore under such circumstances important to have 
the salt and ore introduced at the same time. The 
most perfect application of the salt is undoubtedly 
when both ore and salt are crushed together in the 
battery. 

But a point of great importance is the time when 
the salt should be added, if other objects are in 
view. If the salt is added together with the ore, 
or after the sulphur is expelled and sulphates are 
formed, in every instance the base metals will take 
up their share of the chlorine, and therefore more 
salt will be required. But as the most of the 



30 ROASTING OF OEES. 

chlorides are volatile, the salt is the means of get- 
ting rid of a great deal of the metals during the 
roasting, which in some instances is not very de- 
sirable. For instance, if a great deal of antimony 
and copper is in the ore, more or less chloride of 
silver will escape; sometimes, however, only a 
small percentage. Treating the ore with salt 
from the beginning, or adding it two hours after 
the beginning, the result is the same. 

A different result is obtained if the salt is added 
after all the base metals are desulphurized and ox- 
idized. Some base metals, as antimony and 
arsenic, will be volatilized and thus gotten rid of, 
but not in so large a proportion as if chloridized. 
Iron and copper remain entirely in the ore, while 
both are volatile as chlorides. The roasting must 
be continued at a light red heat till all sulphates 
are decomposed and the metals oxidized. Apply- 
ing the salt after the dead-roasting, the effect dif- 
fers from the above so far, that the base metal 
oxides now are not chloridized, or only to a small 
extent, while the silver alone (some of which appears 
to be changed to a metallic state, the most, how- 
ever, remaining as a sulphate) will be chloridized. 
But in order to effect this chlorination, from 1 to 2 
per cent, of green vitriol must be added in order 
to accomplish the decomposition of all the salt. 
The copper is lost with the tailings unless smelted, 
or extracted by diluted sulphuric acid. 



ROASTING OF ORES. 31 



Permanent Stirring not Essential. 

20. In roasting the ore with salt, a continual 
stirring to the end of the process is not a necessary 
condition for obtaining a good result. This de- 
pends partly on the time and partly on the nature 
of the ore. As long as the ore is not uniformly 
heated, a diligent stirring is important. The ore 
in the corners is too often neglected while the sul- 
phur is burning, and the exposure of a fresh sur- 
face to the oxygen of the air requires also constant 
work; but as" soon as the smell of the chlorine is 
perceptible, the stirring can be carried on at inter- 
vals of from eight to ten minutes. The chlorine 
which is evolved in the mass ( 23) has better oppor- 
tunity to act on the metals than if constantly stirred, 
whereby more chlorine escapes up the chimney 
without producing any effect. This was proved by 
a comparison of the work of two furnaces. A re- 
volving furnace had a speed great enough to let 
the ore drop constantly through the flame and air, 
while the common furnace was managed by only 
one man, and stirred at intervals. Mr. Atwood 
found 15 per cent, less chloride of silver in the 
roasted ore from the revolving furnace. The blame 
is not with the revolving furnace, but with the 
speed. It proves, however, that, being constantly 
exposed to the air, the chlorine escapes with less 
effect than in the common furnace, where the ore is 
allowed to rest for ten or fifteen minutes, and the 



ovol\oo! hlorim-, boin^ 1 in rontai-i \\ ith tin 1 part 
\vhilo ) 'M-omyh th<> mass, is ponuit t ol lo 

form oomhinatious. O : llara's imvhanioal fun: 

Oiirh t!u> otv is romparatixolv but littlo stinvil. 

6 -M to IM poroont. of ohloritlo of siUor. V 
miximv of i>ro, sa\NbiNt atul ^all, fornunl into 
bricks and oulriiu'tl. ^^ho^\*^l ;!u> silver as a rhlorulo 
thfou v >l> tlu \vliolo mass, \vlurt\ as a niatioi- of 
<-ou-s ( > ( (lu> insulo olio! iu>t riMiu 1 into ilinvt cotitart 
\vitl air. (\>nstant slu>v'liu^ is tu^n^sarv \\iili 
; a natmv. as it \\*Mil,l bako if 



Signs of a good Chloridizing Boasting. 



31. A J>XHH! chloritli/.iiijr roasting should 

'.H> JHM- rojit. t>f tlu silvor ronvrrtoil intt> i-hlo- 

;l\(M- % aiul sli - possible of l>ftse 

mHal fllorils. To asiMM'tnin tlu^ amount of 

rivlt* of silvor at tl>o riul of tho roasting, it is 

uwko two Vbvut oiuMuiuro aiul a- 

half is talvtMi tuit i>f tlu^ t'untavv anvl alKn\<\l to 
i-oil. T\\v 0110 halt' oinu-o a^ u\l out, 

ami ono ^No. H pvjvuvil for tho tire assay rts usual. 
Tho *lhor half oinu :utiXHhu-v 

t'ullv into a small tiltor in a j^lass fvuuiol. Tho til- 
torin?:- papor must pn^joot about ono im-h alx>Y6 
tlu^ QM V solution of livposulphi'r U is 

thon pourod ovor tho oro in tho tiltor, ami th 
oo\\tiniu\l as lon^r as a procipita :iud on 

' :tiim of sulphiolo of sodium to tho til- 



riNu OF OIIIM. Il.'l 

lrrod li-|iinl Thin is Ix'.sl, tried inn. rlejin -h 
l.iilic. I f, :i.|'l.er lillf i in;; ir liM.i-liin ;; , I In- :i< l 1 1 1 i< >n 
of Hlllpllii I" !' '<< xlilllll if) I lie lr.|,r|| dor,; no| pro 
dlicr ;i, preripil;|,(.c, or only :i \rry ;,||"lil olir, ; ;o 
I h.i! I li- liiplid !!,:; iilllir 1 Mill v :i III Mr d:n I-.IT color 
wilJioiil, IOMIII" il", ped'eel, I r;ui;ip;irenr\ , Ilir ;i , ;i\ 

I lr;i,c||('d \\illl U.'irill \V,'d-e|- ,'i.lid |, he Idler 1,'iKeiioTr, 
|)lll Jlllo ,'l porrr|;|||| (||;.|| o| I I !. r \r ,:,r|, ;|lnl dllcil, 

l>\ :i|)|>l\ in", lu-.'ii \\illi ;in nJcoliol |;i,ni|). Tin- lillrr 
I.MI lie rnnovril :iinl liiimril ,<i,|io\(> Ilir :,.!in|.|r 
\Vlirn dr\ , I 1 1- .1 :lir ; o|' I lir lillrr ;ilid I I ir : :.i 1 1 1 pi r 
Jtl'C HllX(ul like Ilir ollirr linll olllirr, ;IIK| l>o|!i <( n 
ribloM pliiccd in Ilir :i;.:;;i\ lui n;icr. 

I I I , ,ll|l|H, , ,| ! I, ,1 I I,, ,',, ,|| |Mll c,| | |n ,,,:,, I, , | ,,, , I 

uinl. rl:il.''ii \vli'>n it r. ll,.,ip : lil llml il, clilorillil.tidll i I licitl !\ 

lililhll.Ml . olllr|AVJH(< I- Ml IIIIK'll Hlllplllll. '.I ; ,U, , UMilM I,',, 
l.il . II I'lll II < lllc.i li|<>. SIlMllM l! IH K clllll.il l<> .!.<.) I illl I II. 

111 iiil . 'I pin. chlni [(in M| il >. i I. ii in. (1, nl, mi; i n." 'l'ii in;; 

Ilir I. M. Illl". II. M. llM!l;i. t Vll, "I lull illl Illllll!) IMU'll, Hlldlllll ll 

in.-iih'. I ..i Ilii . pin |,., ... |,,i|! nn ,,ini' r i . ' . i"li. .1 .nil Im I h. 
ir. mil lirr n:. .;i\ . \IP,I|I. i n.'il ..in,,, inn:. I !>. |, ncli. <l \\illi 

ll'il \\llll I, l\ '\\ Illdl lln- .lll|.li:il, III lillviT III (lllll.nlvcd, Mini 

III. Mm. I : :iiii|>l<- l . h. .il. -I wild lr,|... nl|.liil.' <>| ,., l( |ii, in il.- 
Mi Tllx 1 1 ill M > Vi'. < '"in | in l Hi;; Ilir rrillllhi ill Ilir I III. . 11 ,.i \ ',, 1 1 
I. r;i ,ll y 1 1 ill III I lldW III 1 1. 'I I ill' I lir i il I" I 1 1. 1 I 11,1111 III II I III' Illl VI 'I Wll.ll 
I II IN' , I Illlo II ill) .lull-, Mini )|i> W III lir 1 1 Mill) II, ('I I It trifle. 

Ms |i" .nlpliili <>l Hodn l.itlli/ftd 11, nl' which f i \ 

; ',. , ! ' ,11,1,1,1 

I 1 '.. i '.nil. hi. !. n| !;.,<( I;,,!,,, h .. in i .iiiln, in 

.1 h, in. -II, 111 n. cnicililc, n.ml wli.'ii hijiii'l i nl i ...In''. 

.mi.-. . nl u!|, I, MI ( I.. MM I., n. , ill ml. i \..l .. in .in ill |in cm, 

"i in-. Inn.' | MI HP iMHlni" up I,, .nl. ijtlf I'oti I rm mi 

MM n [ilnlc HIM! <h . ...l\ . in \\.il.r II l.il.. . i.i'vriiil l.'.ii . I,. 
1'iirc III' ''.lull. .n ;i|,|.. ,n . |n i li'clly ch'.ir, '.h.,uni" n \.||..w - 

i. I'll- iiltnvc He l.l.i.-l Itc.lllii, nl \\ IP n .|i;i.wii ..II II I i i. ;nly 
IMC ii-i... 

Tiir operation l;,kr.. Irs:; limr if Ilir Ii y po: ;i 1 1 pi, i Te 
:.olnl.on i , u:.r.| in ;, |,,,| :,l;,|r. All clilondr of 

.ii\ri .mil ulao Hulphate of wilvor, if prevent, in 
2* 



34 ROASTING OF ORES. 

dissolved by the hyposulphite and carried off, be- 
side the base metal chlorides. The two assays, 
when ready, are compared, and the difference 
shows the silver which was converted into a chlo- 
ride. For instance, if No. 1 assayed 83 ounces per 
ton, and No. 2 from the filter 4 ounces, the differ- 
ence, 79, is that part which became chloridized. 
That is, 

83 : 79 = 100 : x 95 per cent. 

22. Toward the end of the roasting very little, 
if any, sulphate of silver will be found in the ore; 
but even if a small percentage of it should remain, 
it may, for the purpose of amalgamation or ex- 
traction, be considered equal to chloride of silver;, 
for as soon as it dissolves in water, it becomes a 
chloride, precipitated by the salt, of which a part 
is always yet found undecomposed in the ore. To 
obtain a general idea of the amount of soluble base 
metal chlorides and sulphates, it is sufficient to 
put a small sample of about half an ounce on the 
filter as before, and to leach it with hot water. 
The leach obtained is tried again with the sulphide 
of sodium. A thick precipitate shows that a large 
amount of soluble chlorides is in the roasted ore. 
If a reaction of copper is expressly desired, ammo- 
nia should be used in place of the sulphide of so- 
dium. In presence of 'much iron the precipitate 
will appear brown. This precipitate must smell 
strongly of ammonia. If copper is present, a clear 



ROASTING OF ORES. 35 

blue liquid will be seen above the iron precipitate 
after some time; or the whole maybe brought on a 
filter to separate the liquid from the precipitate. 

Means of Destroying Base Metal Chlorides. 

23. It is very difficult to get rid of all the base 
chlorides. They are formed under the action of 
chlorine and hydrochloric acid. The most of the 
metal chlorides are volatile, and a part is carried 
off through the chimney. Another part of the 
chlorides gives off some of its chlorine, whereby 
sulphates, undecomposed sulphurets, antimonates, 
arsenates and free oxides, are chloridized. Chlo- 
rides which are disposed to transfer chlorine to 
other metals in combination with sulphur or 
arsenic, are: the proto-chloride of iron and of cop- 
per, "the chlorides of zinc, lead and cobalt. When 
in this way the most of the metals are chloridized, 
the base metals, principally iron and copper, are 
losing their chlorine gradually, being first converted 
into sub-chlorides and then into oxides. The 
roasting for this purpose must continue with in- 
creased heat, even when the chlorination of the 
silver is finished. At an increased heat, the base 
metal chlorides lose their chlorine, while the chlo- 
ride of silver remains undecomposed, unless a very 
high temperature should be applied. This process 
requires a long time, consequently also more fuel. 
The decomposition of these chlorides is greatly 
assisted by the use of 5 to 6 per cent, of carbonate 



36 BOASTING OF ORES. 

of lime in a pulverized condition. Lime does not 
attack the chloride of silver, but it is not advisable 
to take too much of it, as it would interfere to 
some degree with the amalgamation. The pulver- 
ized lime rock must be charged toward the end of 
the roasting. First, two per cent, is introduced by 
means of a scoop, the whole well mixed, and then 
examined either with sulphide of sodium ( 22) or 
in the following way : 

A small portion of the roasted ore is taken in a 
porcelain cup or glass, and mixed with some water 
by means of a piece of iron with a clean metallic 
surface. If the iron appears coated red with cop- 
per, some more lime must be added. In place o2 
iron, especially if no copper, but some other base 
metal is present, some quicksilver is mixed with 
the sample. In the presence of base metal chlo- 
rides, the quicksilver is coated immediately with a 
black skin. 

When endeavoring to expel the base metals by 
heat, the loss of silver, in presence of much anti- 
mony, lead and copper, should be investigated 
very carefully. Under certain circumstances it 
is not uncommon to find a loss of even 50 per 
cent, of the silver, if the chloridizing roasting is 
carried on at a high heat for a great length of time. 
The loss increases with the duration of roasting 
and with the degree of temperature. When such 
ore is under treatment, it is necessary to take sam- 
ples during the roasting, and to examine the same 
for the amount of chloride of silver, and also for its 



BOASTING OF ORES. 37 

loss, and to stop roasting when the highest per- 
centage of chloride of silver is obtained, without 
reference to the condition of base metals. 

Steam Decomposes Base Metal Chlorides 
Effectively. 

The formation of base metal chlorides can be 
avoided by a proper but more expensive roasting 
( 33). It requires, first, an oxidizing roasting, 
with the application of steam. This roasting must 
continue until all the metals are desulphurized and 
converted into oxides. When this is accomplished, 
salt and green vitriol are added, and the roasting 
continued until all the silver is chloridized. 

There is also a very good way of getting out a 
great deal of the base chlorides of the ore before 
the silver is amalgamated or extracted, by leaching 
the ore with hot water ( 77). 

Application of Steam in Roasting. 

24. The application of steam in roasting is ad- 
vantageous, for the reason that hydrochloric acid 
is created by the decomposition of chlorides, which 
in turn decomposes the sulphurets. The Irydrogen 
decomposes also the chloride of silver, which, upon 
being reduced to metallic condition by its affinity 
for chlorine, in turn decomposes the hydrochloric 
acid. The silver may thus change repeatedly from 
a metallic condition to a chloride, while the base 



38 ROASTING OF ORES. 

metal chlorides are reduced to oxides, and in that 
state do not interfere with the amalgamation or 
precipitation. The application of steam, however, 
requires a great deal more fuel during the roast- 
ing. Taking the moisture of the fuel into consid- 
eration, there is no roasting done without steam, 
although with a limited quantity. 

Silver Ore, Containing Lead, Unfit for a 
Chloridizing Roasting. 

25. Lead has a bad influence in amalgamation 
and precipitation, and even in the roasting itself, 
causing a baking of the ore at the slightest undue 
rising of the temperature. The chloride of lead 
amalgamates easily, especially in iron pans. Ores 
with 8 to 15 per cent, of lead still allow of a success- 
ful roasting. A part of the formed chloride of lead 
escapes in gaseous form, another part is reduced 
by degrees to oxy-chloride of lead. This latter 
combination goes mostly into the amalgam. If 
there is more lead in the ore than 15 per cent., it 
gives sometimes, according to its nature, as much 
as 85 per cent, of silver, and the retorted amalgam 
is submitted to cupellation in order to separate the 
lead. In Hungary (Offenbanya), black copper, con- 
taining, besides the silver, 10 per cent, of lead, 
is subjected to a chloridizing roasting. The 
pulverized copper is mixed with 12 per cent, of 
salt, 1 per cent, of green vitriol, and 3 percent, of 
saltpetre. The saltpetre oxidizes the lead to a sul- 



ROASTING OF ORES. 39 

phate, which is not affected in the subsequent 
amalgamation in barrels. 

Difference in Roasting Ore for Pan Amal- 
gamation, as compared with that for 
other Modes of Extraction. 

26. The roasting of silver ores, if imperfect, 
will give a better result by amalgamating in an 
iron pan, than in wooden barrels or by precipi- 
tation. This is due to the better decomposition of 
sulphates and undecomposed sulphurets under the 
grinding muller. The roasting for pan amalga- 
mation is therefore less delicate. However, when 
once at work, it is always better to do the roasting 
properly; but it is not necessary to sift the ore 
after roasting in order to separate the lumps from 
the mass, as is done with the barrel amalgamation, 
except to prevent nails from coming into the pan. 
The formation of such lumps, however, must be 
avoided as much as possible. Imperfect roasting 
in the presence of base metals, gives in pans always 
a low fineness of bullion. 

Examples of Local and Different Roast- 
ings. 

27. Boasting of Silver Ores in Freiberg (Saxony) . 
The amalgamation process, and consequently this 
kind of roasting, was given up long ago at Frei- 
berg ; but the method of roasting as performed 



4:0 ROASTING OF ORES. 

there is nevertheless very interesting. The ore 
subjected to roasting consisted of silver glance, 
brittle silver ore, ruby silver, metallic silver, 
fahl ore, bournonite, zincblende, sulphide of anti- 
mony, iron, copper, and nickel pyrites, and of 
gangue, viz: quartz, calc, brown, heavy, and fluor 
spar. It contained from sixty to one hundred 
ounces of silver per ton. 

The dry crushed ore was first spread on a plat- 
form ; on this a layer of damp ore, from the wet 
concentration, was laid, and then 10 per cent, of 
salt. This order was repeated from six to eight times. 
The stratified mass was mixed thoroughly by means 
of shovels and a coarse sieve. This mixture con- 
tained from 9 to 10 per cent, of moisture. For 
this reason, after a charge of 450 to 500 pounds 
was introduced through a hole in the roof, the fire 
was kept very low, in order to dry it at a dark red 
heat, and the ore was diligently raked by two men, 
working alternately. As soon as the decrepitation 
of the salt ceased, the ore was ridged from the 
bridge toward the flue, through the middle of the 
furnace, and the formed lumps broken up by iron 
hammers attached to long handles. After this was 
done, the heat was increased, whereby the ore, 
under constant stirring, assumed a red hot con- 
dition, and the sulphur commenced to burn quite 
lively. This stage was reached in two hours from 
the beginning. The desulphurization commences 
with the burning of the sulphur, creating a tem- 
perature sufficiently high to continue the roasting 



ROASTING OF ORES. 41 

without fuel for some time. Fumes are evolved, 
consisting of steam, antimony, sulphurous acid, 
arsenic, etc. This desulphurization takes again two 
hours, the workmen all the time raking and chang- 
ing the hotter part of the ore at the bridge with 
the cooler at the flue. The temperature is now 
raised to a light red heat, the ore increases in 
volume, emitting chlorides of metals, chlorine, hy- 
drochloric acid, etc. The formation of the chlo- 
rides progresses rapidly, and is finished in three- 
quarters of an hour. The charge is then drawn 
out. A too long roasting would not give an 
equally good result, as some silver might be de- 
composed to the metallic state, which is not so 
readily amalgamated as the chloride. 

In what Condition the Metals are after 
Roasting. 

28. After roasting, the silver is found almost 
entirely converted into a chloride, but a small part 
may remain as a sulphate ; antimonate of silver 
also is formed. The iron is converted into an oxide, 
some into sulphate and ar senate and basic chlo- 
ride. The copper appears also oxidized, with some 
sub-chloride and less chloride. Lead remains prin- 
cipally as a sulphate and basic chloride. Zinc is 
oxidized. Antimony is found as anthnonates com- 
bined with oxide of antimony, and with other basic 
oxides. Nickel and cobalt remain as oxides, chlo- 
rides and arsenates, and the arsenic is found as an 



42 ROASTING OF ORES. 

arsenate combined with other nietals. Besides 
these metal combinations, there is uiidecomposed 
salt, sulphates of soda, of lime, etc. The charge 
loses about 10 per cent, of its weight, of which a 
part is regained from the dust chambers. 

Antimonate of silver is considered the cause of 
the loss of "ilver in roasting, as it is not decom- 
posed in barrels, and consequently not amalga- 
mated. There have been no experiments made as 
yet to determine whether aiitimonate of silver is 
decomposed in iron pans during the amalgamation 
or not. Probably it is not, as the natural com- 
bination of antimony, lead and silver, cannot be 
amalgamated in pans. 

39. Roasting of Argentiferous Copper Ores. At 
Arivaca (Arizona) the silver ores from the Heinzel- 
mann mine consisted principally of silver copper 
glance (stromeyerite) with 51 per cent, of silver; 
fahl ore with from 2 to 15 per cent, silver, contain- 
ing also some quicksilver ; zincblende, galena and 
other decomposed argentiferous copper ores. On 
an average the ore contained from $150 to $200 per 
ton, and from 10 to 15 per cent, of copper. After 
crushing, the ore was spread on a platform covered 
with 8 per cent, of salt, and mixed thoroughly by 
means of shovels. Eight hundred pounds of it, as 
a charge, were introduced through the roof of the 
furnace (which was constructed entirely of adobe), 
spread on the hearth, and, at a dark red heat, 
stirred for two hours, at the end of which time the 



ROASTING OF OEES. 43 

flame was colored an intense greenish-blue, and 
considerable fumes were emitted. The raking con- 
tinued for an hour and a half more at an increased 
heat, and during this time the ore was moved three 
times from the bridge to the flue and back. A 
sample taken from the furnace at this time, put on 
a canvas filter, wet with salt solution and leached 
with a hot concentrated solution of salt, gave a 
clear liquid, which, diluted with water, showed a 
strong white precipitate of chloride of silver, mixed 
with antimony and lead ; but the quicksilver, 
treated with the same sample of ore and water, 
was cut and blackened to a high degree. For this 
reason, from 5 to 6 percent, of pulverized lime was 
thrown into the furnace by means of a scoop, as 
much as possible over the whole surface of the ore, 
and then raked and stirred diligently in order to 
finish the mixing in the shortest time. After four 
hours from the beginning, the temperature was 
raised to a light red heat for half an hour, and the 
roasting was finished. There were yet a great 
many base metal chlorides in the ore, but as metal- 
lic copper was used in the barrels, the silver turned 
out always over 900 fine. The loss of silver was 
12.5 to 13 per cent. 

30. Roasting of Copper Matt. In smelting ar- 
gentiferous copper ores, the process is sometimes 
regulated to produce a sulphide of copper, contain- 
ing silver and base metals, as antimony, arsenic, 
zinc, iron, etc. This sulphide of copper, or copper 



44 ROASTING OF ORES. 

matt, was roasted formerly and smelted again to 
produce black copper ; that is, impure metallic 
copper. For the purpose of extracting the silver 
therefrom, the copper was melted together with a 
certain percentage of lead, and the latter, with the 
silver, extracted by liquation and cupelled. The 
remaining copper contained still some silver and 
lead, and the process was a very lengthy one be- 
fore finished. To avoid the liquation, the copper 
matt was treated by amalgamation, and the silver 
extracted at once. For this purpose the matt was 
crushed and sifted, and the coarse part ground. 

Of this powdered matt, 300 pounds are charged 
in a double furnace, of which the upper hearth 
prepares the ore by a moderate roasting, while the 
lower one finishes the operation at a higher tem- 
perature. In each of the hearth-departments the 
matt is treated for two hours and a half. Silver 
and the other metals are first converted into sul- 
phates and then mostly decomposed to oxides, but 
the silver remains for the greatest part in metallic 
condition. The matt is drawn out, mixed with 8 
per cent of salt and 12 per cent, of lime, and with 
salt water into a paste, which is allowed to rest for 
twelve or fourteen hours. The paste is then dried, 
powdered between rollers, and again roasted two 
hours and a half. During this process, samples 
are taken and mixed with water and a few drops of 
mercury. If this appears coated bluish, it proves 
the presence of metallic salts, and some more lime 
must be added ; if after this the quicksilver re- 



ROASTING OF OKES. 45 

mains perfectly white, parting, however, in many 
minute globules, it proves that too much of the 
lime was used, and in this case some of the first 
roasted matt is added. 

The purpose of wetting the roasted ore, as above 
described, is the formation of chloride of silver. 
As there are always sulphates of the metals pres- 
ent after the first roasting, they decompose the 
salt, and the chlorine acts on the metallic silver. 
This process is not perfectly finished, and therefore 
the second roasting. 

31. Roasting of Black Copper. The black cop- 
per obtained from smelting (in Schmoellnitz, Hun- 
gary) contains from 110 to 150 ounces of silver per 
ton, and 85 to 89 per cent, of copper. To pulver- 
ize this it must be made red hot in a reverberatory 
furnace and crushed while red hot. The powder 
must be sifted and then ground fine. The pulver- 
ized metal is then mixed with 7 to 9 per cent, of 
salt, and roasted in the usual way for six to six 
and a half hours, in charges of 400 pounds each. 
No green vitriol is added for the purpose of de- 
composing the salt ; and as there is not more, than 
from \ to 1 per cent, of sulphur in the black cop- 
per, the salt decomposes through direct action on 
the copper. First, chloride and subchloride of 
copper are formed. The copper chloride transfers 
chlorine to the metallic silver, and is reduced to a 
sub-chloride. 

In other places iron pyrites are added to the 



46 ROASTING OF ORES. 

black copper, by which the chlorination is pro- 
moted. At Oriklowa (Banat) 5 per cent, of iron 
pyrites and 12 per cent, salt are used, roasting 
twelve hours. The loss of silver is 7 percent., and 
of copper 3 per cent., during. the roasting. The 
expense of roasting is $7.30 per ton. 

32. Roasting of Silver Ore at Flint, Idaho Ter- 
ritory, in O'Hara's Mechanical Furnace. The ore 
from the Rising Star mine, at Flint, contains ar- 
gentiferous fahl ore, miargyrite, ruby silver, zinc- 
blende, galena, iron pyrites and sulphide of anti- 
mony. On an average the ore paid between $90 
and $100 per ton, containing some gold. The 
gangue is quartz. It is crushed through sieves with 
forty holes to the inch, together with 5 per cent, 
of salt. The furnace ( 57, Fig. 8) is charged con- 
tinually by machinery at one end, a. The ore is 
moved by degrees forward, and arriving at the first 
fire-place, c, commences to disengage sulphur. 
Between this fire-place and the second, which is on 
the opposite side, between c and d, the chlorination 
begins at an increased heat. The flame shows 
partly a blue color, originating from chloride of 
copper, and white fumes are also evolved. Be- 
tween the second and the third fire-place, d, the 
chlorination is finished at a light red heat. From 
the cooling hearth, e, the roasted ore is continually 
discharged on the dump,/*. It takes six hours be- 
fore the ore from the feeding place, a, arrives at 
the dump. Although not more than 5 per cent, of 



ROASTING OF ORES. 47 

salt is added, the roasted ore contains about 90 per 
cent, of the silver converted into a chloride. The 
gases, containing free chlorine and chloride com- 
binations, emitting chlorine (16), being in contact 
with the surface of the ore while passing over it 
for a space of eighty feet, have a chloridizing influ- 
ence on it, replacing thus a certain amount of salt. 
These three furnaces roasted twenty tons of ore 
in twenty-four hours. The expenses were as fol- 
lows : 

For wood, five cords, at $5 $ 25 00 

For four men at the furnaces, at $4 16 00 

For two men bringing in wood, etc., at $4. . 8 00 

For one man as watch, in the night 4 00 

For blacksmith work 5 00 

For 2,000 pounds of salt, at 8c 160 00 

Total expense $218 00 

or $10.90 per ton. The capacity of the three fur- 
naces is calculated for more than twenty tons. 
Each one could easily treat ten tons of the Rising 
Star ore in twenty-four hours. The roasted ore 
was treated by amalgamation in pans, applying the 
" leaching process." 

Taking moderate prices of salt and labor into cal- 
culation, instead of the " Flint Tariff," the roasting 
expense in O'Hara's furnaces would not exceed 
$5.50 per ton, provided three furnaces were at work. 

33. Roasting of Silver Ores for the Patera Pro- 
cess. The ores treated by Patera's process ( 71) 
at Joachim sthal are remarkable for the numerous 
mineral species occurring in the ore. Among these 



48 



ROASTING OF ORES. 



may be mentioned silver, lead, different compounds 
of copper, bismuth, iron, uranium, nickel, cobalt, 
etc., sulphur, arsenic and antimony. Before the 
introduction of Patera's process, the extraction of 
silver, on account of so many base metals, was very 
difficult. The success of Patera was not so much 
in adopting hyposulphites, as proposed by Percy 
and Hauch, but in his modification of the roasting 
process, by which only the silver was converted 
into a chloride. 

The pulverized ore is placed in the furnace, in 
charges of from four to five hundred pounds. 
First quite a moderate heat is applied, and gradu- 
ally increased, but not so much as to induce clot- 
ting. As soon as the ore appears red hot, steam is 
admitted, with about four pounds pressure to the 
square inch. As the steam consumes heat, more 
fuel must be used to keep up a red heat. The ore 
must be constantly raked during the whole period 
of this roasting. It takes about four hours to 
finish this process, after which the ore is drawn 
and permitted to cool. The iron appears now as 
an oxide also the copper ; some sulphate of cop- 
per is also present, and the silver is principally in 
the state of a sulphate. 

The oxidized ore is now ground finer, mixed with 
from 5 to 12 per cent, of common salt, and, at the 
same time, with 2 to 3 per cent, of calcined green 
vitriol. This mixture is spread upon the hearth in 
the furnace, and subjected to a second, now chlo- 
ridizing, roasting. It takes about an hour before a 



ROASTING OF ORES. 49 

red heat is reached. Steam is then introduced, as 
above, under continuous stirring. The fire is gradu- 
ally increased, and the roasting finished within from 
five to eight hours, according to the value of the 
ore. There are condensing chambers for catching 
volatile metals and ore dust. They are of impor- 
tance if rich ore is treated, and without this con- 
trivance several per cent, of silver would be lost. 
When finished, the ore is drawn and allowed to lie 
undisturbed for some time, after it has been moist- 
ened. In this condition the chlorination con- 
tinues. The application of steam causes nearly 
twice the consumption of fuel, but it has been shown 
that, by the steam, hydrochloric acid is formed, 
whereby arsenic and antimony are expelled, and at 
the same time the chlorination of the silver greatly 
favored. Over one ton of coal and one-half cord 
of wood are consumed for each ton of ore roasted 
in this way. 

34. Roasting for Augustin's Process. To this 
process principally matt is subjected. The process 
requires a chloridizing roasting. The ore, if not 
itself rich in sulphurets, is mixed with iron py- 
rites, slag, lime, etc., and smelted. The molten 
sulphide of iron takes up the silver and deposits 
itself below the slag on the bottom of the furnace. 
The silver is thus separated from the earthy part 
and concentrated in the matt. Argentiferous cop- 
per ores are likewise smelted for the purpose of 
obtaining argentiferous copper matt. The matt is 
3 



50 ROASTING OF ORES. 

then finely pulverized, and 400 pounds of it are in- 
troduced into the furnace. The roasting goes on now 
in the usual way, by starting with a moderate tem- 
perature, gradually increasing it, exposing every 
portion of the ore to the intense heat by frequent 
stirring, etc. At the end of eight hours the roast- 
ing is generally completed, the matt looks dark 
and earthy, and no fumes of sulphurous acid can 
be perceived. The roasted stuff is now drawn out 
and permitted to cool. After this the matt powder 
is ground finer, sifted or bolted, and given back to 
the furnace. 

Four hundred pounds are mixed with 5 per cent, 
of salt. Sulphates are present, but oxides of 
metals predominate in the mixture. The formation 
of the chlorides commences immediately, and the 
roasting is concluded after one or two hours. The 
temperature in this second Coasting is kept low, as 
the smelting of the chloride of silver must be pre- 
ventedfor when melted the chloride of silver 
dissolves with more difficulty in a salt solution. In 
other places, after the first roasting, the mass is 
not taken out and re-ground, but, when finished? 
only a portion is drawn, mixed with from 1 to 6 
per cent, of salt, according to the purity of the 
matt, charged again, mixed with the balance of ore 
in the furnace, and roasted for one-half to three- 
quarters of an hour. 

35. Boasting of Silver Ores for the Chlorination 
Process ( 74). It is not absolutely necessary for 



BOASTING OF ORES. 51 

this process to have the ore roasted with salt, but 
it has been found that, on account of different 
earths, an addition of 1 or 2 per cent, of salt pro- 
duces a better result. This process extracts cop- 
per, gold and silver, each of which is obtained 
separately; but it makes a difference in roasting, 
whether the copper is intended to be saved or not, 
as in many localities the copper is at present of no 
value, or the old iron for precipitation is too ex- 
pensive. 

The ore is crushed dry through a sieve of -forty 
holes to the running inch. Some ore allows also 
thirty holes to the inch. In case the ore contains 
so much clay or talc that no leaching is admis- 
sible, the ore is crushed wet, separated from slime, 
and dried . Eight hundred to one thousand pounds 
are charged, and, according to the quality of the 
ore, the heat raised quickly or slowly to a bright 
red heat. This is an oxidizing roasting, conse- 
quently much stirring is required. Ores with but 
few sulphurets appear sufficiently well roasted 
after three hours ; other ores containing an abun- 
dance of sulphurets, take from five to six hours and 
more before all sulphur, arsenic and antimony are 
expelled ; and the arsenates and antimonates 
formed by the two last are not volatile. When de- 
sulphurized, 1 or 2 per cent, of salt is thrown in 
the furnace and mixed with the ore as intimately 
as possible. Three-quarters of an hour after the 
addition of the salt, and with only a moderate heat, 



52 ROASTING OF ORES. 

the roasting' is finished. Ores, not rich in sul- 
phurets, may be mixed with salt when charged. 

If it is intended to extract the copper also, this 
must be transformed into chloride of copper. To 
accomplish this, two things must be observed, 
first, no oxide of copper should be formed during 
the roasting ; and second, more salt must be used. 
Stoichiometrically, each pound of copper requires 
1.84 pounds of salt to form a chloride, provided 
all chlorine is taken up by the copper ; but as 
this is not the case, as a great deal of chlorine is 
also absorbed by other metals, etc., it follows that 
at least two pounds of salt, if not more, must be 
taken for each pound of copper in the ore. For 
most localities such a quantity of salt could not be 
used on account of the difficulty in obtaining it. 
It may be mentioned here that if the brine, remain- 
ing after the copper has been precipitated by iron, 
should be condensed by evaporation, exposing it to 
the heat of the sun, which might be practicable on 
the Pacific Coast in the dry season, the condensed 
salt, consisting of chloride of iron, could be added 
to the ore as a chloridizer, whereby a considerable 
percentage of salt would be saved. 

In roasting with salt with reference to extracting 
the copper, the ore is first roasted for itself at a 
low temperature, so as not to decompose the sul- 
phates by a bright red heat, but long enough to 
decompose all sulphurets.. "When this is accom- 
plished, the salt is introduced and the roasting 
finished in one-half to three-quarters of an hour 
thereafter. 



ROASTING OF ORES. 53 

36. Roasting of Silver Ores in a Long Furnace at 
San Martial, Sonora, Mexico. For the purpose of 
roasting silver ores for the Solving and Precipi- 
tation Process, there are several long furnaces 
(thirty feet long) built up by Mr. O. Hofmann, in 
Sonora. Using long furnaces is 'found a great 
economy in every respect. A great advantage re- 
sults also from moving the ore, at intervals of from 
one to three hours, from one hearth to the other. 
By this means it is impossible, even with careless 
roasters, to find raw ore in the finished charge, as 
would happen under such circumstances if the cor- 
ners of single roasting furnaces w r ere not very care- 
fully attended to during the roasting. 

The furnace at San Marcial is sixty feet long. 
The plan was made by Mr. Graff, who superin- 
tended the reduction works. It is a level hearth 
sixty feet long, representing six furnaces, each ten 
feet long, parted only by the projecting wall in- 
side, as shown by v, Figs. 6 and 7 ( 49). There 
are six working doors on one side, and, on account 
of the length, an auxiliary fire-place is placed at the 
back side, before the last two hearths. Each hearth 
contains 800 pounds of ore, and is attended by two 
Yaque Indians at a time, stirring alternately. 
These twelve Indians perform all the work about 
the furnace, carry the wood from the adjoining- 
yard, split what is too thick, carry out the ashes, 
etc. The ore on the first hearth, nearest to the 
fire-place, has always a light red heat, which de- 
creases with the distance from the bridge, so 



54 ROASTING OF ORES. 

that the fifth appears quite dark if not assisted 
by the second fire-place. After stirring one hour 
on all six hearths, the charge from the first is 
drawn out through a door in the rear, on a large , 
smooth platform, arid immediately spread by means 
of shovels in a layer one or one-and-a-half inches 
thick, so as to have it cool enough for transporta- 
tion to the sifting apparatus after the lapse of one 
hour. As soon as the hearth is cleared, the ore on 
the second hearth is moved over to the first, then 
that on the third to the second, and so on, till 
the sixth hearth remains empty and is charged 
through the funnel in the roof with a new charge. 
It will be seen that 800 pounds of ore are drawn 
out every hour, and that each charge is exposed to 
the fire for six hours. It is thus evident that, 
being moved six times from one hearth to the other, 
the ore arrives perfectly prepared to the finishing 
heat. After roasting and sifting, the ore is amal- 
gamated in pans, but as it contains some carbonate 
and sulphuret of lead the amalgam is charged with 
base metals, so much that refining by cupellatioii 
is necessary. From 8 to 10 per cent, of salt is 
added and mixed with the ore before it is charged. 
Preparations are made now to introduce the solv- 
ing and precipitation process, if successful on that 
kind of ore. 

The above arrangement, employing so many 
hands, is considered by Mr. Graff a local neces- 
sity; the Indians are cheap, but not very attentive, 
laborers. 



ROASTING OF ORES. 55 

According to an analysis made by Mr. Graft", 
the roasted ore from single furnaces (treating $100 
ore) contained 5 per cent, less of chloride of silver 
than that from a long one. Long roasting furnaces 
are especially adapted for roasting sulphurets con- 
taining gold. Concentrated sulphurets, or ore con- 
taining an abundance of sulphurets, allow the use 
of a very long furnace, with only one fire-place, 
on account of the heat created by the burning 
sulphur. 

The roasting expenses at San Marcial, with the 
furnace sixty feet long, are as follows : 

2i men, day and night, at 50c $12 00 

2 cords of wood at $3 6 00 

8 per cent, of salt=l,536 Its. at 2c 30 72 

Kepairs, etc 3 00 

Total expense on 9.8 tons $51 72 

or $5 27 per ton. 

A furnace thirty feet long, with the same kind of 
laborers, 800-pound charges, drawing every two 
hours, that is, 4.8 tons in 24 Ijours, shows the 
following expenses : 

8 roasters at 50c $ 4 00 

1% cords of wood at $3 4 50 

8 per cent, of salt at 2%c 19 20 

Other expenses 2 00 

Total $29 70 

or $6 18 per ton. 

37. Roasting in Stetefeldt's furnace at Reno, 
Nev. The mechanical part of the roasting itself, 



50 ROASTING OF ORES. 

in tliis furnace, is the simplest of all, and also the 
shortest. The finely pulverized ore, mixed with 
salt, is sifted continuously by a mechanical ar- 
rangement into a shaft. This shaft, a, (Fig. 9, 
58), is about twenty-five feet high, and heated by 
two fire-places provided with grates. The ore, 
falling through the heated shaft, undergoes chlo- 
rination, a process requiring only a few seconds. 
After the roasted ore has accumulated on the bot- 
tom of the shaft to the amount of about 1,000 
pounds, it may be drawn out. The amount of 
salt needed for chlorination, varies according to 
the ore; generally about 6 per cent, or 120 pounds 
to the ton, is taken ; or even less, especially in 
treating poor ores, when half of that amount may 
be sufficient in most cases. A furnace having a 
capacity of from fifteen to twenty tons in twenty- 
four hours, consumes from two' to three cords of 
wood. In twenty-four hours there are employed : 
Two men attending the feeding and conveying ma- 
chinery, three firemen, and three men to draw and 
cool the roasted ore. As the latter three have 
time enough to carry the ore to the pans, only half 
of their time should be charged to the roasting 
expenses. According to these figures, the total 
expense of roasting, in Reno, is not more than 

For labor of 6% men at $3 $19 50 

For wood, 2% cords at $6 15 00 

Salt, 1,800 pounds at l%c 27 00 



Total expense on 15 tons $61 50 

or $4 10 per ton. From 88 to 92 per cent, of the 



ROASTING OF ORES. 57 

silver contained in the ore is converted into a chlo- 
ride. Of the dust in the dust-chambers, the silver 
was found in the state of a chloride up to 96 per 
cent. [See Scientific Press, 1869, p. 377]. 

It is evident that with an improper treatment of 
the fire, by using too much or too little fuel, a less 
favorable result would be obtained. In the first 
place, if the temperature is kept too high, a part 
of the chloride of silver is reduced to a metallic 
state, which, for the purpose of amalgamation, is 
not so very injurious (8); but the metallic silver 
is a total loss with the solving process. On the 
other hand, if there is not sufficient heat, some of 
the sulphurets may remain undecomposed. In 
either case the responsibility is with those in 
charge of the furnace; but there is nothing easier 
than to keep up a proper and uniform heat in 
Stetefeldt's furnace, there being no other hand- 
work about it, and all the attention of the fireman 
being directed to this single point. 

Stetefeldt's furnace has been compared some- 
times with Gerstenhoefer's shelf furnace. In ref- 
erence to this the Engineering and Mining Journal 
says : 

' ' Since the discovery and exploration of the num- 
berless mineral deposits in the Western States 
and Territories, no branch in metallurgy has re- 
ceived so much attention as the process of roast- 
ing ores of all descriptions. One can hardly look 
over a file of Mining Journals, or newspapers from 
some mining district, without finding descriptions 
3* * 



58 ROASTING OF OKES. 

of new devices for roasting ores, all of .which 
claim to surpass everything else in this line 
which was known before. The devices are as 
strange as they are many, and much time and 
money has been w r asted to test impracticable in- 
ventions. Indeed, the high expense which the 
roasting in the old reverberatory furnace entails, 
was a strong inducement to invent some cheaper, 
and at the same time more effective, method. 
This is especially of importance where silver ores 
are found, which require a chloridizing roasting 
preparatory to their amalgamation. In such cases 
the expense of roasting is frequently more than 
one-half of the total expense of reduction, and 
consequently low-grade ores cannot be worked 
with a profit. But in spite of the necessity to 
adopt some improved and more economical pro- 
cess of roasting, it has been extremely difficult to 
introduce two inventions, which are based upon 
the most simple and rational principles so sim- 
ple, indeed, that it seems impossible to simplify 
them any more. We speak of the Gerstenhoefer, 
or Terrace furnace, first introduced about six years 
ago at Freiberg, and the Stetefeldt furnace, in- 
vented three years ago at Austin, Nevada, but first 
introduced for regular working at the mill of the 
Nevada Silver Mining Company, near Reno, Ne- 
vada, in October of last year. The nature of these 
inventions can be expressed as follows : 

Gerstenhoefer discovered that sulphurets are 
completely roasted or oxjjlized if they fall against 



ROASTING OF ORES. 59 

a current of hot air rising in a shaft, which is 
filled with shelves, so as to check and retard the 
fall of the ore particles at certain intervals. 

Stetefeldt discovered that silver ores, no matter 
in what combination the silver occurs, mixed with 
salt, are completely chloridized if they fall against 
a current of hot air rising in a shaft with no ob- 
structions whatever to check or retard the fall of 
the ore particles. 

It is a matter of course that in both cases a cer- 
tain degree of fineness is required for the ore to 
be treated, and that a much coarser material can 
be successfully roasted in the Gerstenhoefer fur- 
nace than in Stetefeldt's. 

We do not intend to enter here into a detailed 
description of the Gerstenhoefer furnace, since 
that invention has been frequently laid before the 
public in several mining papers; but we will 
merely compare it with the Stetefeldt furnace, 
and point out the distinctions of the two inven- 
tions. 

As a cheap chloridizing roasting is a vital ques- 
tion for the industry of silver mining in this coun- 
try, it is evident that Stetefeldt's discovery far 
surpasses that of Gerstenhoefer in importance. 
But the question arises, whether the former, as 
constructed by Gerstenhoefer, cannot be used as 
well for chloridizing as desulphurizing roasting? 
"We answer, no. In the Gerstenhoefer furnace only 
such ores can be successfully treated, which, at a 
red heat during roasting, have no tendency to 
sinter or stick together. 



60 ROASTING OF ORES. 

But the small particles of a charge of ore mixed 
with salt are exactly in such a condition while 
roasting, as to have the greatest possible inclina- 
tion to sinter and adhere to the shelves. Tfiey 
would thus soon obstruct the whole shaft, and 
prevent any further w r ork. This has been demon- 
strated by actual experiments on a working scale. 

From the foregoing, it is apparent that the ap- 
plication of the Gerstenhoefer furnace, even for 
desulphurizing purposes, is very limited, and that 
certain classes of ore must be entirely excluded 
from it. This is especially the case with galena 
ores, which are the most expensive to roast in re- 
verberatory furnaces. 

In Stetefeldt's opinion, the shelves in the Ger- 
stenhoefer furnace are perfectly superfluous, and all 
ores, even galena, can be desulphurized by drop- 
ping them through a plain shaft heated by fire- 
places below, if they are reduced to a sufficient de- 
gree of fineness. The escape of unroasted dust 
from the shaft is of no consequence, as a separate 
fire-place is constructed for the roasting of these 
suspended particles in the Stetefeldt furnace. 
Furthermore, the feeding machinery of the Stete- 
feldt furnace is based upon a principle entirely 
differing from that used with the Gerstenhoefer 
furnace. 

That a furnace without shelves is cheaper and 
easier to construct, more durable, less liable to get 
out of order, and that it requires less labor and 
skill to run it, must be conceded by everybody. 



ROASTING OF ORES. 61 

Much difficulty was experienced to provide suit- 
able feeding machinery for the Stetefeldt furnace. 
Gerstenhoefer's apparatus, consisting of fluted 
rollers, which force the ore through slits in the top 
of the furnace, would not answer at. all. The ore 
fell down in lumps, and arrived at the bottom of 
the shaft almost raw. The reason for this be- 
havior is simply the tendency of the particles of 
all finely-pulverized mineral substances to adhere 
to each other if a slightly compressed mass of them 
falls through the air. It is, therefore, necessary 
to introduce the ore pulp so finely divided, that all 
the particles can be penetrated by the heat within 
the short time of their fall through the shaft. To 
feed the pulp with a blower, as it is done in Keith's 
desulphurizing furnace, was not considered desir- 
able for the following reasons : 

1. The fall of the ore would be accelerated. 

2. The draft of the fire-places would be im- 
peded by the downward current of the air from the 
blower. 

3. The formation of dust would be considerably 
increased. 

The feeding machinery in its present shape can 
be briefly described as follows: 

A hollow cast iron frame, kept cool by a small 
stream of water, rests on top of the furnace. 
In this frame is inserted a cast iron grate, which 
is covered by a punched screen of Russia iron, 
No. for wet crushing, of the trade. Close to the 
punched screen moves, inside of the hopper, a 



62 ROASTING OF ORES. 

coarse wire screen, No. 3 of the trade, which is 
fastened to a frame. The frame has flanches rest- 
ing upon adjustable friction rollers outside of the 
hopper, and receives its motion from a crank, with 
If-inch eccentricity. To avoid the motion of a 
stratum of pulp with the coarse screen, a number 
of thin iron blades are so arranged across the hop- 
per that their lower edges reach close down to the 
coarse screen and keep the pulp in place. When 
the crank is set in motion, the meshes of the 
coarse wire screen cut through the pulp and 
drive it through the openings of the punched 
screen. In this way the ore is introduced in a 
continuous stream into the furnace. The motion 
of the crank-shaft was variably tried in Reno, at 
from thirty to seventy revolutions per minute." 

The very satisfactory result of roasting silver 
ores in Stetefeldt's furnaces at Reno, induced the 
Manhattan Silver Mining Co. to adopt the same in 
their mill at Austin, Nevada. Mr. Stetefeldt altered 
the plan of the Reno furnace somewhat, omitting 
a dust-chamber between the furnace and the verti- 
cal flue. Another important improvement is the 
application of gas generators ( 58) in place of the 
arrangement for the use of firewood. The gen- 
erators are fed with charcoal. After the proportion 
of air (which is admitted through a separate flue 
and is necessary for the burning of the carbonic- 
oxide gas) has been regulated, the uniform heating 
of the furnace does not depend on the skill of the 
fireman. There are three gas generators at the 



ROASTING OF ORES. 63 

furnace, two communicating with the furnace 
shaft, a, of Fig. 9, 58, and one with the vertical 
flue, b. One separate gas generator could supply 
all three entrances, but Mr. Stetefeldt prefers to 
have them separate, partly on account of dispens- 
ing with iron gas pipes and other inconveniences 
connected with a general generator. The choice 
of firewood or gas arrangement depends on local 
circumstances the comparative price of wood and 
charcoal, etc. 

The furnace, as represented in Fig. 9, 58, is 
calculated to roast from twenty-five to thirty tons 
of ore in twenty-four hours, at a cost of from $5 to 
$6 per ton, while the expenses in usual reverbe- 
ratory roasting furnaces at Austin amount to $12 
or $14 per ton. 

Stetefeldt 's furnace is of vital importance, espe- 
cially for poor ores requiring roasting. Consider- 
ing the temporary repairs of the arch and floor of 
the reverberatory furnaces, the amount of black- 
smith's work on hoes and rakes, besides what was 
stated above, the superiority of Stetefeldt's furnace 
over the reverberatory and other mechanical fur- 
naces is obvious. 

It may be remarked that wherever the use of the 
solving process ( 60) appears admissible on silver 
ores, this, in connection with Stetefeldt's roasting, 
will allow the most economical extraction of silver, 
even from very rebellious ores. The baking of the 
ore during the chlorination, in the presence of 
lead, cannot take place in Stetefeldt's furnace, and 
it is therefore very probable that a higher amount 



64 ROASTING OF ORES. 

of lead will be less injurious than in any other 
roasting process. 

Considering the old, or rather the usual 
theory deduced from the roasting process in com- 
mon reverberatoiy furnaces, that sulphates must 
be formed before the salt can be decomposed, and 
not till then will the chlorination begin, it would 
seem that for these chemical reactions more time is 
required than a few seconds ; but this is not the 
case. As soon as ore and salt enter the furnace, 
each sulphuret particle ignites in the glowing at- 
mosphere, evolving at the same time sulphur, which 
in presence of the oxygen of the atmospheric air, 
coming undecomposed through the grates, is 
turned into sulphurous acid and the metal into an 
oxide, or in part directly info a chloride. The 
sulphurous acid in contact with the ore particles 
and oxygen becomes sulphuric acid. The temper- 
ature is nearly from the start too high to permit 
the formation of sulphates, so that the sulphuric 
acid turns its force on the red hot salt particles, 
setting the chlorine free. All these reactions are 
performed, instantaneously. Steam, emanating 
from the fuel, is also amongst the gases, conse- 
quently the creation of hydrochloric acid must 
ensue. The whole space in the furnace is filled 
with glowing gases of chlorine, hydrochloric acid, 
sulphurous and sulphuric acid, oxygen, steam, vol- 
atile base metal chlorides, etc. all of them acting, 
decomposing and composing, on the sulphurets 
with great vigor. The chlorine decomposes the 



ROASTING OF ORES. 65 

sulpliurets directly, forming chloride of metals and 
chloride of sulphur ; it attacks decomposing^ also 
oxides and sulphates, if present. The hydrochloric 
acid performs the same office. Also metallic silver, 
if it should occur in the ore, would combine with 
the chlorine. The sulphuric acid, besides decom- 
posing the salt, oxidizes partly the sulphurets di- 
rectly, etc. 

Considering now an ore particle in a red hot 
condition attacked simultaneously by all these 
gases while falling, the final chloridizing result is 
inevitable. The finer the ore particles are, the 
more perfect the chlorination ; but even if some 
coarser parts (to a certain degree) should reach the 
bottom not thoroughly chloridized, this would be 
finished in the pile, as the chlorination and evolu- 
tion of chlorine gas continues in the red hot accu- 
mulation on the bottom of the furnace. 

38. Chloridizing Roasting of Silver Ore contain- 
ing Gold. Generally the ore containing silver and 
gold is roasted with salt, converting thereby the 
silver into a chloride, while the gold remains in a 
metallic condition. This mode of roasting is quite 
satisfactory for the subsequent amalgamation in 
pans. But if those metals are intended to be ex- 
tracted by a solving process, where no amalga- 
mation takes place, the gold also must be converted 
into a chloride while roasting. By roasting ores in 
which gold and silver is present with salt, chlo- 
ride of gold is formed, according to Plattner ; but 



66 ROASTING OF ORES. 

before the ore becomes red hot, the gold loses a 
part of its chlorine, is reduced to a sub-chloride, 
and, at a little higher degree of heat, to metallic 
gold. 

To form, chloride of gold by way of roasting, a 
better result is obtained in the furnace if the ore is 
roasted first without salt till the smell of sulphur 
is no longer perceptible ; and then, after it has 
cooled down to a low temperature, the salt is added 
and the whole stirred for some time. A suitable 
form for a furnace would be a long hearth furnace 
( 49), altered in such a way that the second hearth 
should be ten or twelve inches below the first ; the 
arch, however, should continue in a straight line. 
By this means the space is widened and the tem- 
perature brought down to the proper degree. The 
ore is charged on the first hearth near the bridge, 
and roasted in the usual way, oxidizing the sul- 
phurets. When this is effected, the ore is shifted 
over in the lower furnace, and the upper charged 
again. As soon as the roasted ore assumes a dark 
red heat, the salt is introduced and raked for two 
or three hours. According to Roeszner, a combi- 
nation of gold oxide of soda and chloride of sodium 
(Au 2 O 3 Na Cl) is formed. It is not soluble in 
water, and but slightly in a salt solution, and can- 
not be amalgamated, but it is soluble in hyposul- 
phite of soda or lime. V. Lill and others consider 
the gold in the state of a sub-chloride (Au Cl). Hot 
water cannot be used for the purpose of leaching 
out base metals, as the chloride of gold would be 
decomposed. 



ROASTING OF ORES. 67 



Proper Roasting Charges. 

39. The endeavor to perform the roasting in 
the most economical way, leads many operators 
astray, since they lose sight of the great impor- 
tance of the chemical reactions, which, as the main 
object, have to be" considered first. Everything 
has its limit, and so the quantity of ore to be taken 
for one charge. The European fashion (charging 
from four to five hundred pounds) could not be 
well adopted in the United States, while to place 
2,000 pounds at once in a furnace might do for a 
mere desulphurization, but is decidedly too much 
for both the chloridizing and the oxidizing roasting 
of silver ores, if good results, by one or the other 
mode of extraction, are to be obtained. Eight 
hundred pounds, or at least not over 1,000 pounds, 
is a permissible quantity in a properly constructed 
furnace, and with careful handling. 

B. Oxidizing Roasting. 

40. The purpose of the oxidizing roasting is 
either to expel volatile substances which are com- 
bined with the metals (as sulphur or arsenic), or to 
expel volatile metals which are considered ob- 
noxious to further treatment of silver ores (as anti- 
mony, lead, zinc, etc). The oxygen has a large 
share in this transaction, and combines with the 
volatile substances, as well as with the metals. 



68 ROASTING OF ORES. 

Some of the combinations with the ox} r gen become 
volatile as, for instance, sulphurous, arsenous 
and antirnonious acids, lead and zinc oxides, etc. 
Other combinations again are not volatile, as the 
formed sulphates, arsenates and antiinonates. 
Some of these latter compounds can be disengaged 
by an increased heat, as the sulphates of iron and 
of copper, whereby the sulphuric acid escapes, 
while the remaining metal turns into an oxide. 
Others cannot be decomposed by an increased heat, 
or an increased heat is considered injurious for 
other reasons ; and in this case such combinations 
may be decomposed by an addition of charcoal 
powder, saw-dust, or the application of hydrogen. 
The sulphuric acid is reduced by the carbon to 
sulphurous acid, and goes off, and so also the 
arsenic and antimony. The carbon deprives the 
sulphate or arsenate of a part of its oxygen, and 
escapes as carbonic acid. 

Changes of the most common Metals while 
Roasting. 

41. Iron pyrites and copper sulphurets suffer 
different changes, based on the action and reaction 
of the oxygen, sulphurous acid, sub-oxides and sul- 
phurets, before a complete change is effected into 
sulphates of iron and copper. Raising the heat, 
the sulphuric acid is driven out, and iron oxide 
and copper oxide remain unchanged. If salt is 
present, in place of the oxides first, chlorides are 



ROASTING OF ORES. 69 

formed, whereby both became volatile. Galena, or 
sulphuret of lead, at a low temperature, turns by de- 
grees into oxide and sulphate of lead, the first 
being volatile. At a higher heat the sulphate re- 
mains unchanged and cannot be decomposed into 
an oxide. In presence of salt the greatest part of 
the lead becomes a chloride, which is volatile, but a 
part of it loses some chlorine, and thus being re- 
duced to a basic chloride, is no longer volatile. 
Antimony is volatile as an oxide, but there is also 
antimonic acid formed, which combines with other 
metal oxides to antimonates, which are not volatile 
and not decomposed by increased heat. As a chlo- 
ride the antimony is very volatile. Zincblende, or 
sulphur el of zinc, requires a long roasting. Oxide 
of zinc is formed besides the sulphate. The sul- 
phate of zinc, under stronger heat, is reduced to a 
basic sulphate, which can be decomposed to an 
oxide, but only at a white heat. As a chloride it 
volatilizes. 

What Process Requires Oxidizing Roast- 
ing. 

/ 

42. The oxidizing roasting is in use partly as a 
preparatory treatment for a chloridizing roasting. 
It is independent only for Ziervogel's process and 
for the chlorination process ; that is, so far as con- 
cerns the extraction of silver by precipitation. For 
amalgamation of silrer ore no oxidizing roasting is 
suitable; but it is important with the smelting pro- 



70 ROASTING OF ORES. 

cesses, and also in extracting gold from gold ores 
principally from sulphurets (iron pyrites). For 
this purpose long furnaces ( 49) are the most 
suitable. 

The main point in the roasting for Ziervogel's 
process is the creation of a sulphate of silver, and 
the oxidation of the base metals as far as possible. 
As the decomposition of sulphates of different 
metals depends on different degrees of tempera- 
ture, such roasting appears of a very delicate na- 
ture. To this process principally argentiferous 
copper matt is subjected. 

43. Boasting of Copper Matt. When pulverized 
until fine enough to pass through a sieve t>f thirty- 
three holes to the running inch, the mass is intro- 
duced into the furnace and spread out by means of 
rakes. The matt inclines much to clotting. For 
this reason a very moderate temperature is applied, 
more for drying than for roasting. The matt is 
left -quiet for about fifteen minutes, after which the 
stirring is commenced and continued without stop- 
ping for an hour. During this time many lumps 
are formed, which the roaster tries to crush to 
powder. Near the working door the stuff is ex- 
posed to a draft of fresh air, in consequence of 
which the roasting on that place progresses more 
rapidly than it does further back. This makes a 
shifting of the stuff necessary after one hour's 
roasting. The other roaster now takes the rake 
and stirs the matt again for an hour, doing the 
work precisely as the first roaster did. The roast, 



ROASTING OF ORES. 71 

ers change in this way every hour for five and one- 
half to five and three-quarters hours. This roasting 
is performed on the upper hearth of the double 
furnace. Twenty -five pounds of coal dust are 
mixed with the matt, causing an ignition and 
emission of gases, and the whole mass is transferred 
to the lower hearth through a hole in the bottom. 
The upper hearth is now charged with 500 pounds 
of matt anew. 

The sulphur commences to burn after a raking of 
three-quarters of an hour, and the mass increases 
in volume when the hearth is covered about four 
inches with matt. During the roasting all metals 
are converted into sulphates, of which, toward the 
end of the operation, iron and zinc vitriol are de- 
composed, leaving those metals as oxides. Copper, 
nickel and cobalt remain in the rtate of sulphates. 

The lower hearth is in a light red hot condition 
when the matt falls in from the upper hearth. To 
prevent the rapid burning of the admixed coal 
dust, and the clotting of the mass, a vigorous stir- 
ring for an hour, with closed dampers, is strictly 
observed. The stuff is now shifted and then the 
damper opened. There now follows a sharp oxi- 
dizing roasting, with free access of air, for one hour 
and a half. By means of the air current, the 
roasting mass is cooled down so far that it appears 
quite dark. To see the progress of the roasting, a 
sample is taken out and examined, either on a porce- 
lain dish or on a filter with cold water. The 
leach must appear of a clear blue color, and an 



72 



ROASTING OF ORES. 



addition of salt solution must give some white pre- 
cipitate, proving the beginning of the formation of 
sulphate of silver. If the nitrate shows a greenish- 
blue color, the presence of sulphate of iron is ap- 
parent, and in this case the oxidizing roasting must 
continue. 

The purpose of the addition of coal dust is the 
reduction of sulphates to basic salts, whereby sul- 
phurous acid is emitted. With the opening of 
the damper the oxidation progresses, the sulphate 
of iron is decomposed almost entirely, the sub-oxide 
of copper turns into oxide, and when the oxidizing 
roasting is finished, the mass contains mostly 
oxides, but also basic salts. There are copper, 
iron and zinc oxides, sulphates of copper and zinc, 
while the silver as yet consists principally of an 
undecornposed sulphide. The next stage of roast- 
ing at an increased temperature is the last one. It 
is directed toward the sulplmrization of the silver 
and complete oxidation of the base metals. It 
takes two hours and a half to accomplish this re- 
sult, under continuous raking and increasing the 
temperature to a light red heat. Samples are 
taken again as before, and examined in the same 
way. The leach must appear only of a bluish tint, 
and on adding salt solution, a heavy precipitate 
must fall, caused by chloride of silver. The whole 
roasting period on the lower hearth, as on the 
upper, takes from five and one-half to five and 
three-quarters hours. 

The formation of the sulphate of silver in the last 



ROASTING OF ORES. 73 

period at an increased heat, is due to the sulphuric 
acid in gaseous form, emanating from the sulphate 
of copper. It attacks the sulphide of silver, and 
combines with it to a sulphate. Ninety-two per 
cent, of the silver is extracted after this roasting. 
If in the last period the feeding with fuel should 
be carelessly performed, so as to give a smoky 
flame, some copper oxide will be reduced to sub- 
oxide, and this will precipitate metallic silver while 
leaching, causing a loss. If the roasting should 
not continue long enough, some sulphide remains 
undecomposed ; and, on the other hand, if the 
roasting should last too long, a part of the sulphate 
of silver would be decomposed to metallic silver 
and could not be leached out. These circum- 
stances show that this kind of roasting demands a 
great deal of attention, in order to obtain a perfect 
result. The temperature on the lower hearth in 
the beginning is 500 to 550 degrees Centigrade ; it 
sinks then to 425, and rises again at the end of 
the operation to 770 degrees . 

44. Boasting of Gold Ores. The gold is gen- 
erally found in a free state as metallic gold. In 
this state it is easily extracted by proper amalga- 
mation. Often, however, the gold is combined 
with other substances, so that amalgamation is of 
no avail unless the gold is set free by roasting. 
Iron pyrites and arsenical pyrites are the principal 
ores containing the gold in a condition unfit for 
direct amalgamation. Also telluride of gold must 
4 



74 ROASTING OF ORES. 

be subjected to roasting before amalgamated or 
chlorinated, but this mineral is not often found. 

The roasting of sulphurets and arseniurets is 
very simple, all that is required being a perfect, 
dead roasting ; that is, expulsion of all sulphur 
and arsenic ; but this process takes generally more 
time than a chloridizing roasting. After the fur- 
nace has been heated for some hours, the sulphu- 
rets are introduced into it and spread over the 
roasting hearth, which is generally about twelve 
feet by twelve, and is capable of receiving one ton. 
One man is sufficient to attend a single furnace, 
but a long one requires two men. A single furnace 
commences with a low heat, sufficient to start the 
self -burning of the sulphurets, by which so much 
heat is created as for several hours to require but 
very little fuel. Nearly half of the sulphur is ex- 
pelled with this low heat. On exposing a fresh 
surface of the mass by stirring, the burning of the 
sulphur with a bluish flame can be seen distinctly. 
The hoe is principally used . for stirring. It must 
be as light as possible, seven to eight feet long, if 
prepared to work from both sides of the furnace. 
The stirring is performed at intervals of ten to fif- 
teen minutes, but not longer ; and wherever the 
circumstances permit two roasters to be employed, 
the time of roasting will be shortened. Oxidizing 
roasting requires more stirring than the chlo- 
ridizing. 

In proportion as the oxidation of the sulphurets 
draws nearer to the end, the temperature decreases, 



ROASTING OF ORES. 75 

and it is then necessary to use more fuel to keep 
the mass at a good red heat. It takes from twenty 
to forty hours before the roasting of one charge in 
a single furnace may be considered finished. If, 
in throwing up sulphurets in the furnace, by means 
of a shovel or hoe, many brilliant sparks appear, 
this denotes that the roasting is not finished, but 
must be continued till this appearance ceases. 

In a long furnace (Figs. 7, 8, 49) the hearth 
near the bridge is always kept at a bright heat. 
One man attends to the ore on the first hearth, and 
thje other two or four hearths can be managed by 
a second. In moving the ore from one hearth to 
the other, or in drawing the charge from the' fin- 
ishing hearth, these two men assist each other. 
The finishing hearth receives the ore already desul- 
phurized to a great extent, and containing only a 
small part of undecomposed sulphurets, but more 
of sulphates. With a lively heat and active stirring 
at intervals, all base metals ought to be converted 
into oxides after ten or twelve hours. 

An addition of from thirty to fifty pounds of salt 
to the ton of sulphurets at the end of roasting, two 
or three hours before the discharge, is not injurious 
to the subsequent chlorination, but it increases 
uselessly the expense if mixed with such sulphurets, 
when there is no necessity for it. In many in- 
stances, however, especially where the sulphurets 
contain lime, calcspar, talc or heavy spar, a chlo- 
ridizing roasting is necessary, if it is intended to 
extract the gold by chlorination. One hundred 



76 ROASTING OF OEES. 

pounds of salt are sometimes required for one ton 
of sulphurets. 

Sulphurets containing gold can be brought into 
a soluble condition by means of roasting, accord- 
ing to 38, so that no chlorination is required 
after roasting. 

Roasting Furnaces. 

45. Roasting not only requires much care, but 
it is also an expensive operation. For this reason 
the choice of the right kind of furnaces is of very 
great importance, and so much the more as a per- 
fect and economical extraction of silver depends 
principally on the result of roasting. The chlori- 
dizing roasting is known to be the most suitable 
way for the subsequent extraction of silver in what- 
ever way it may be performed, by amalgamation or 
solving ; consequently those furnaces in which the 
ore particles are exposed to the action of chlorine 
and other chloridizing gases to the most advantage, 
must be considered the best. The old style of 
furnace was four to six feet wide and ten feet long, 
and in them a small part of the ore was exposed to 
the greatest heat near the bridge. The gases 
evolved were carried along by the draft, being in 
contact with the surface of the ore for a length of 
ten feet while passing over it ; but on account of 
the narrowness of the hearth, the ore at the bridge 
had to be changed often with the cooler part at the 
flue. 



ROASTING OF ORES. 77 

The next step in improvement was the adoption 
of wicler hearths, even wider than long. The heat 
was more uniform and the result better. In both 
kinds of furnaces the chlorination of the metals de- 
pends principally on the chlorine developed in the 
mass of the ore while passing through it ; but once 
above the surface, the chlorine and volatile chlo- 
ride metals have little chance to transmit their 
chlorine to the ore ( 23), and this only through 
the chlorination period. During two or three hours 
of each charge, when desulphurization and sul- 
phatization are going on, this must be performed 
by the oxygen of the air, while, if chlorides were 
present from the beginning, sulphurets, sulphates 
and oxides would have been partly decomposed di- 
rectly by the chlorine, whereby time and a certain 
percentage of salt are saved. 

In this respect a great advantage is gained by 
the introduction of "long furnaces" ( 49), in 
which a continual formation of chlorides on the 
finishing hearth near the bridge is going on, vola- 
tile chlorides and free chlorine being evolved, 
which, on their way to the flue, are constantly in 
contact with the ore for a space of thirty or fifty 
feet in length. These furnaces show a great 
economy in fuel, labor and salt, and the roasted 
ore contains a better percentage of chloride of 
silver ( 36). 

Another most important improvement in the way 
of chloridizing roasting is found in the Stetefeldt 
furnace ( 58), where all ore particles are involved 



78 ROASTING OF ORES. 

in chloridizing gases under very favorable circum- 
stances. The roasting is cheap, and from twenty 
to twenty-five tons of ore are roasted in twenty- 
four hours more than ever accomplished in any 
other furnace. 

The roasting furnaces do not require a white heat; 
hence common bricks can be used ; but it is nev- 
ertheless advantageous if the fire-place above the 
grates is built of fire bricks. In new or unpopu- 
lated districts even unburned bricks or adobe may 
be used ; they stand just as well as burned bricks 
of the same material, except on the floor of the fur- 
nace, which is worked out in two or three months. 
Hard bricks are the best material for the hearth- 
floor, placed edgeways (four-inch), with as little 
clay between as possible, and laid carefully and 
well fitting, so as to form a level and smooth sur- 
face. All parts exposed to heat must be built with 
loam or clay, not with mortar. Many masons have 
the custom of laying three heights of bricks so that 
the eight-inch wall is formed by two rows length- 
ways, and only the fourth height is put crossways. 
It is a quick work and may answer for buildings, 
but should not be allowed with furnaces where the 
expansive heat must be considered, especially in 
the fire-place. Each alternate row of bricks must 
be laid crossways to the preceding ; also, adjusting 
the wall with the hammer, to make it perpendicular 
and square, after several bricks are laid, is injurious. 
The outside appearance of a furnace is of minor im- 
portance, and the mason must, contrary to his gen- 



BOASTING OF ORES. 79 

eral idea, pay the most attention to the solid and par- 
ticular work inside. The distance of the arch from 
the hearth is from eighteen to twenty inches in the 
highest point, not far from the bridge ; in a long 
furnace, however, the roof of the first hearth can 
be higher from the floor by four to five inches, ac- 
cording to the length. An eight-inch thickness of 
the arch is sufficient, and the bricks laid with the 
eight-inch side perpendicular form a more dura- 
ble arch than one of twelve inch thickness com- 
posed of eight and four-inch sides of the bricks. 
The furnace must be secured against expansion by 
grappling-irons of cast iron tightened with iron 
rods from five to six-eighths of an inch in diameter. 
The rods placed over the length of the furnace are 
stronger one inch in diameter. In place of iron 
grapplings, also wooden posts, six by eight inches, 
are used, tied by iron rods on the top. The lower 
ends are generally put in the ground, but it is 
preferable to use rods on both ends. In case of 
need, even the rods are replaced by timber. For 
the passage of the rods square holes must be pro- 
vided, in the masonry; also for the escape of 
dampness such passages are necessary at different 
points, especially if the whole block consists of 
masonry. The floor of the hearth should be three 
feet and a half above the ground ; if lower, it is 
inconvenient for the roaster. 

There are two principal classes of furnaces such 
as are managed by hand and such as employ ma- 
chinery. For the first class mostly reverberatory 



80 ROASTING OF ORES. 

furnaces are in use. The second class comprises 
reverberatory, cylindrical and vertical furnaces. 

A. Roasting Furnaces Managed by 
Handwork. 

46. Beverberatory Furnaces. Reyerberatory 

furnace is the name applied to all horizontal hearth 
furnaces provided with grates and fire-place on one 
side, and a flue connected with a chimney on the 
other. The draft here is created by the chimney 
instead of by bellows, as in blast furnaces ; there- 
fore only such fuel is used which gives a flame, and 
consequently no charcoal, coke or anthracite is 
serviceable unless in a gas reverberatory furnace, 
where gas (carbonic oxide) is produced from char- 
coal or other fuel sometimes alsa by the aid of 
compressed air and burned. The reverberatory 
roasting furnaces are constructed in various ways. 
There are single furnaces, with but one hearth, and 
double furnaces, with two hearths, one above the 
other. Sometimes above the second hearth there 
is a third one for the purpose of drying the charge. 
Long furnaces also are coming into use. 

47. A Single Roasting Furnace is represented by 
Fig. 3, showing the section, and Fig. 4, the ground 
plan. The bottom, a, or the hearth, is made of 
the hardest bricks, laid edgewise and as close as 
possible. Some masons lay the bricks flat. This 
mode is cheaper and quicker, but far inferior and 



ROASTING OF ORES. 



81 



less durable than the former way, and requires a 
more carefully prepared foundation. The very 

Fig. 3. 




2345 



best bricks must be selected for the hearth, b 
4* 



82 ROASTING OF ORES. 

shows the discharge hole in front of the hearth. 
It is more convenient to draw the ore toward the 
front hole than to have a door for this purpose be- 
hind, but circumstances may decide fo* such dis- 
charge doors. The flue, e, is in connection with 
the flue-holes, e f , in the arch, as indicated by dotted 
lines in Fig. 4, and is from nine to ten inches in 
diameter. The flue-holes in the arch have the ad- 
vantage that no ore can enter when being stirred, 
as often happens when the flue commences at the 
hearth. The distance between arch and hearth 
near the bridge is twenty to twenty-one inches, and 
near the flue only eight inches. The flue leads 
into the chimney in any suitable direction, either 
directly or through a dust chamber. Often the 
flue is led under the floor (when the chimney is at 
some distance from the furnaces), and is made 
wide enough to serve as a dust chamber say two 
feet wide and three feet high, or wider if several 
furnaces are connected therewith. The chimney 
is from twenty to fifty feet high, and from one and 
one-half to three or four feet square in the clear. 
On the top of the chimney an iron cover, controlled 
by a chain, regulates the draught. This is prac- 
ticable only when but one furnace is attached to 
the chimney, otherwise dampers must be provided 
for each furnace in the flue. The bridge, i, is 
much exposed to injury by fire on one side, and by 
raking on the other ; it is therefore advantageous 
if the upper part, or the whole bridge, can be made 
in two or three parts and of some fire-proof stone, 



ROASTING OF ORES. 83 

sandstone, granite, or some conglomerate, which 
does not burst when heated. The grates, h, are 
twelve to sixteen inches below the top of the bridge, 
eighteen inches wide, and from six to seven feet 
long. The space between the grate-bars is one- 
fourth to one-half of an inch. 

In the roof, nearer to the bridge, is an opening 
four to five inches square, of cast iron, in con- 
nection with a funnel, I, of sheet iron. This fun- 
nel must be large enough to receive one charge of 
the ore. A slide keeps the ore in the funnel. 
The roof must be either eight inches thick, or the 
double length of brick ; that is, sixteen inches. 
Under the hearth there is an arched space, d, into 
which the roasted ore is drawn through the dis- 
charg'e hole, 6, either directly into an iron car or on 
an inclined floor, on which the ore slides from un- 
derneath the furnace. In front this space is shut 
up by brickwork. For the purpose of easy drying- 
it is well to leave open some holes, g, for the escape 
of dampness. It is not necessary to build the 
block under the hearth solidly of bricks. The 
space inside is generally filled up with rubbish of 
bricks and stone. 

The working door, o, is from twenty-five to thirty 
inches wide. In front of it is an iron roller for 
easier handling of the heavy tools. The door is 
eight to nine inches high. The cast iron door- 
frame, p, for the fire-place, is from nine to twelve 
inches square. When completed, the furnace is 
tied by iron rods, n, both ways. The uprights 



84 



ROASTING OF ORES. 



are generally wooden ones, six by six or five by 
eight inches. 

It is very important to dry the furnaces, when 
finished, with a very moderate fire for five or six days, 
day and night. Upon a slow, gradual drying, the 
durability of the arch depends. The furnace must 
be nearly red hot before the first charge of ore is 
introduced. 

48. A Double Boasting Furnace is represented 
Fig. 5. 




in Fig. 5, in longitudinal cross section. The lower 
hearth, a, is nine feet long and ten feet wide. The 



OF THB 

UNIVERSITY 



BOASTING OF ORES. 



roof in the center is eighteen inches, and at the 
flue and bridge fourteen inches above the hearth. 
The fire-place, r, is twenty inches wide, eight feet 
long, and twenty inches from the roof. The flue, 
6, ascends to the upper hearth, c, the working door, 
o, of which is on the back side. In case there should 
be required more heat than is obtainable from 
the lower hearth, there is an auxiliary fire-place, r' . 
The flame goes through the flue, &', into the dust 
chamber, g. This chamber has four cross par- 
titions lengthways, by which the draught is forced 
to take a longer way before it enters the chimney. 
From the upper hearth the ore is drawn through 
the hole-, d, to the lower hearth, as the bridge does 
not permit the flue to be used for this purpose. 
e, e, are canals for the escape of moisture. The two 
hearths can be used separately if needed. In this 
case the flue, b, is closed and another one (not seen 
in the drawing) opened. This second flue commu- 
nicates directly with the dust chamber. 

49. Long Roasting Furnace. This kind of 
roasting furnace, as represented by Fig. 6 in vertical 
section, and Fig. 7 in ground plan, gives much 
satisfaction, as there is not only a great saving of 
fuel effected, but also a greater quantity of ore can 
be roasted in a given time than with a single fur- 
nace. It is only a modification of the double 
furnace, but it seems to be more convenient for 
the roasters. The heat is better utilized, as the flame 
has not to pass through flues between the hearths, 



86 



ROASTING OF ORES. 



and is not broken so often, 
but the moving of ore from 
one hearth to the other is 
more troublesome. There 
are two men employed at a 
time, there being one ton 
and a half to two tons in 
the furnace. The hearths 
are either arranged horizon- 
tally, as the drawings show, 
or only the first one is level; 
the other two are inclined ; 
this facilitates the shifting 
of the ore. Each hearth is 
ten feet long and ten or 
twelve feet wide. After the 
first hearth there is a step of 
four to six inches, partly to 
divide the first from the oth- 
ers, but principally to con- 
tract the space between roof 
and hearth of the other two. 
The ore is fed on the last 
hearth through the sheet 
iron funnel, a, spread equally 
on &', and, according to its 
dampness or the quantity of 
sulphurets contained, stirred 
more or less for one and a 
half to two hours. As 
it is not only inconven- 



ROASTING OF ORES. 



87 




venient, but im- 
possible to have a 
good stirring ef- 
fected at a dis- 
tance of twelve 
-feet, which re- 
quires long and 
heavy tools, there 
are for this reason 
working doors on 
both sides of the 
furnace . The 
roaster uses hoes 
or rakes eight feet 
long, made partly 
of gas pipe, which 
are light and han- 
dy. The working- 
doors are thirty 
inches wide. They 
must all be kept 
closed except when 
the ore is being 
raked, and then it 
is very proper to 
have half of the 
door closed (with 
a piece of sheet 
iron). Sufficient 
air comes in at the 
Hl^ working door of 
the first hearth. 



88 ROASTING OF ORES. 

After one and a half to two hours the ore is re- 
moved to the second hearth, from b to c' and from 
b' to c, and again spread over the whole of c, c' . 
Another charge is introduced onb,b f . The second 
hearth has a better heat than the third one. The 
ore is treated here as before, being raked as often 
as possible. After a lapse of one and a half to two 
hours the ore is moved again to the first hearth, in 
the same way as before ; that is, from c to d' and 
from c' to d. The ore is now exposed to a light 
red heat, by which the chlorination or oxidation 
must be finished in the same time as on the other 
hearth. It is necessary to change here the ore 
from the bridge toward the flue, and reverse once 
during the roasting. When the operation is fin- 
ished, the roasted ore is drawn into iron cars below 
the furnace through the opening, e. When all the 
ore has been removed, the charge on the second 
hearth is transferred to the first, from the third to 
the second, and from the funnel to the third 
hearth, and the process continued as before, so 
that a thousand pounds are drawn out every one 
and a half or every two hours. 

The bridge,/, is fourteen inches high. For the 
purpose of admitting air or steam, a canal can be 
made in it. The fire-place, g, is eighteen inches 
wide and eight to nine feet long, and fifteen inches 
below the top of the bridge. The ash-pit, h, is 
made according to what seems more convenient, 
as represented either in Fig. 6 or in Fig. 3. A deep 
ash-pit is more favorable for the preservation of 



ROASTING OF ORES. 89 

the grates, as they are less heated. Each door is 
provided with an iron roller, i. A furnace of a 
similar description is in operation in La Dura 
(Mexico), roasting refractory silver ores for the 
chlorination process. 

A furnace sixty feet in length, with six hearths, 
as built by Mr. Graff at the San Marcial, has the 
advantage of being capable of roasting from eight 
and one-half to twelve tons of ore in twenty-four 
hours, discharging every hour from eight hundred 
to one thousand pounds, according to the charge. 
In case ore is subjected to roasting which has 
not enough sulphur to create the required heat in 
burning, an additional smaller fire-place must be 
attached on one side, so as to bring the flame into 
the fourth hearth. 

Muffle Furnaces. 

50. A muffle furnace, as the name indicates, is 
a furnace constructed of clay and cast iron in such 
a way as to prevent the flame from coming inside 
of it. The fuel heats the mantle or muffle from 
the outside, so that the ore is not heated directly 
by the burning fuel, but by the glowing muffle. 
The muffle furnaces require, therefore, more fuel 
to obtain a certain degree of heat than ordinary re- 
verberatory furnaces, where the flame comes into 
contact directly with the ore. The use of this fur- 
nace is limited, and applicable in cases where the 
air or the gases of the burning fuel are injurious, 



90 ROASTING OF ORES. 

or where volatile substances from the ore should be 
condensed ; as for instance, sulphur, zinc, arsenic, 
etc. .For roasting silver ores, these furnaces are 
not in use, but they were tried in California in dif- 
ferent ways ; also for desulphurization, adding 
charcoal to the pulverized ore. The experiments, 
however, were not successful, as could have been 
anticipated ( 5, e). 

B. Roasting Furnaces with Mechanical 
Arrangements. 

5 1 . There is a great variety of furnaces wherein 
the costly stirring by hand is replaced by mechan- 
ical apparatus. No mechanical furnace can be 
governed in every part of the roasting process with 
the same facility and precision as is possible in a 
reverberatory furnace with manual labor ; but in 
the latter case the great difficulty in finding good 
reliable roasters, and the heavy expenses connected 
therewith, make a mechanical substitution very de- 
sirable. In one respect, vertically revolving fur- 
naces, in which the ore moves without being 
stirred by mechanical rakes or plows, have the ad- 
vantage, viz : in simplicity. The stirring furnaces 
combine sometimes complicated machinery with a 
general defect, and this is the wear of the shoes or 
plows, not on the lower part alone, but also on the 
sides. Some stirrers are fixed, and consequently 
cannot remedy the wear by sinking ; but none 
have been yet so constructed as to keep their 



ROASTING OF ORES. 91 

original line on the wall side. The consequence 
is, that while the shoe wears away, the ore takes 
its place and hardens there (being exposed to heat 
so long) till new shoes are put in '; and in this case 
all the hard cakes are broken off and mixed with 
the ore. But as new shoes are not put in every 
day, these periodical lumps may be sifted and re- 
turned to the battery. The accumulation of ore in 
some parts of the furnace successively, cannot be 
avoided entirely, even in furnaces where no stirring 
goes on. 

A. Stirring Furnaces. 

52. Eevolving Hearth Furnace. The shape of 
this furnace is circular. There is an iron frame of 
from ten to twelve feet diameter, with sides four- 
teen inches high. The whole is lined with brick, 
the bottom four inches thick. The discharge open- 
ing is on the bottom, extending from the periphery 
toward the center, and is four inches wide and 
three feet three inches long. This opening is shut 
by an iron door, hung on hinges. It is not neces- 
sary to fill this space with brick, which would in- 
terfere with the easy opening ; but the space, after 
the discharge of the ore, must be filled up with 
roasted ore, of which enough is always left in the 
furnace. The bottom is fixed to an upright shaft, 
four inches in diameter, provided with a spurwheel 
at the lower end to impart the motion. This ten 
or twelve-foot bottom is surrounded by a substan- 



92 KOASTINQ OF ORES. 

tial ring wall, as close to the periphery of the bot- 
tom as possible. The bottom is then arched over 
with bricks, leaving the doors through which the 
new shoes are introduced when the old ones wear 
out. There is also a cast iron pipe through the 
center of the furnace, on which the shoes are 
fastened and so arranged that one set plows the 
ore against the center, the other set toward the 
periphery. The pipe is hollow and cooled by a 
continual stream of water. There is also a hole 
four to five inches square in the arch, in connection 
with a funnel, through which the ore is charged 
into the furnace. The distance from the bottom 
to the center of the arch is thirty-one inches. The 
arch is connected on one side with the fire-place, 
six or seven feet long and eighteen or.twenty inches 
wide, and about ten inches below the rim of the 
revolving hearth are the grates. On the opposite 
side is the connection with the flue. 

Such furnaces have the advantage that they carry 
the ore in a circle, so that each part is equally ex- 
posed to the heat near the bridge and to the cooler 
region near the flue. While revolving, the funnel 
is opened and the ore falls on the moving bottom, 
being spread in passing under the stationary stir- 
rers, which are of a plow shape. The roasting 
takes about the same time as in an ordinary fur- 
nace, but requires less fuel, as the furnace is not 
cooled down by air, which enters the common re- 
verberatory furnace through the working door. It 
is important to have the horizontal shaft provided 



ROASTING OF OEES. 93 

with two driving wheels of different size, so that 
about one to two revolutions per minute can be ob- 
tained while roasting, and from six to eight revo- 
lutions while discharging. After the roasting is 
finished, the discharge door on the bottom can be 
opened, while the hearth revolves slowly. In this 
furnace it is an easy matter to arrange the plows 
in such a way that they could be moved every 
second or third day toward tho periphery as much 
as they wear off. In this way the side of the 
hearth can be kept always clear from accumu- 
lation of the ore crust. 

53. A similar furnace is Brunton's revolving 
furnace. The hearth has a low conical shape, the 
highest point being the center. Above this is the 
charging hole in the roof. The hearth is twelve 
feet in diameter, and takes one ton of tin ore at a 
charge. There is a cast iron rake with three-inch 
long prismatic teeth, which are dovetailed and 
so constructed as to be easily replaced. The ore 
comes through a funnel in the center of the revolv- 
ing hearth, and is spread by the stationary rake, 
the position of which is not radial but oblique. 
The hearth is fixed to a solid vertical shaft with 
gearing, by which a slow rotating motion is im- 
parted to the hearth, so that only one revolution is 
made in forty minutes. 

54. Ernst's Rotary Furnace is constructed on 
the same principle as the former two, but differs 



94 ROASTING OF ORES. 

materially in two points. While in the before de- 
scribed revolving hearth furnaces, the diameter can- 
not be increased well beyond twelve or fifteen feet, 
principally for the reason that the roof would stand 
too far off from the bottom, Ernst's furnace can 
be constructed on any given diameter. The other 
point of difference lies in the discharge, which is 
continuous with Ernst's furnace. 

The hearth is a circular iron ring (disk), lined 
with bricks, and revolving on a series of iron roll- 
ers. It is kept in motion "fey means of two gear 
wheels, each three feet in diameter. The speed is 
regulated by a cone pulley, being increased or di- 
minished to conform to the require^ motion of the 
hearth. The ore is charged continuously from the 
battery, through a hopper, by means of elevators. 
The hearth moves constantly, carrying the ore from 
the flue toward the fire-place, and exposing it to 
all the different temperatures of the furnace, thus 
effecting a very uniform heating of the ore. The 
hearth-ring is surrounded by a corresponding wall 
of masonry, leaving a free circular space in the 
center, where the driving machinery is placed. 
There are stationary and movable stirrers. The 
movable paddle-stirrers are connected with the 
gearing under the iron bed of the hearth on its in- 
side periphery. The ore finally arrives at the fire- 
bridge, where it receives its ultimate and highest 
heat, and is then discharged by an apparatus, 
which consists of two chain or rag wheels, by 
which a double endless chain, with inserted plates 



ROASTING OF ORES. 95 

of iron, is carried across the hearth, thus discharg- 
ing the roasted ore. 

This furnace, not as yet in use, promises to do 
good work. In some respects it could be com- 
pared with O'Hara's furnace having a long, nar- 
row hearth ; replacing the endless chain, however, 
by the motion of the hearth, both being self -dis- 
charging. For chloridizing roasting of silver ore, 
the center circle of the ring must have a diameter 
of about twenty-six feet. One revolution should 
be accomplished in five or six hours. It takes 
about twenty stirrers to rake the ore every ten 
minutes, if equally divided. It is, however, more 
proper to arrange the stirrers closer toward the 
feeding place. 

55. Parke's Roasting Furnace, with movable 
stirrers. This is a double furnace, one hearth 
above the other, with a common vertical shaft to 
which the stirrers are fastened. The hearth is 
twelve feet in diameter and rests on an arch, be- 
neath which the rotating motion is transferred to 
the shaft by means of gearing. On one side of the 
lower hearth is the fire-place, whence the flame 
draws over the bridge into the furnace. 

Opposite the bridge is an opening one foot wide 
and four feet long, through which the flame ascends 
to the upper hearth. Both of the hearths have 
two working openings, which are closed by cast- 
iron doors. From the upper hearth the flame 
draws through a flue into the chimney. The shaft 



96 ROASTING OF ORES. 

goes through both hearths and the roof. There 
are two massive arms in both furnaces, with curved 
spikes attached for the purpose of stirring the ore. 
In order to keep the shaft cool, it is hollow, and a 
few holes above the gear permit the cold air to 
draw through the shaft, whereby a constant cool- 
ing is effected. The upper end of the shaft runs 
in a cast iron cross, fixed on the roof of the furnace. 
After the ore on the lower hearth is drawn out 
through the discharge hole at the bottom, the ore 
on the upper hearth, already desulphurized to a 
great extent, is raked toward a similar discharge 
hole, and then transferred to the lower depart- 
ment. The raw ore is charged through the roof 
into the upper part. By means of hoes the ore 
is spread on both hearths, before the shaft is 
allowed to revolve again. 

56. Bruckner's Revolving Furnace. An iron 
cylinder about ten feet long and four feet in diam- 
eter is lined with bricks inside. The cylinder is 
fixed with its long axis in a horizontal position. 
One end communicates with a fire-place, from 
where the flame passes through the revolving cyl- 
inder into the flue at the other end. Inside there 
is an inclined shelf, the position of which is such 
that the ore is being continually shifted from one 
end of the cylinder to the other, as this revolves 
and thereby thoroughly exposed to the flames. In 
this way the ore becomes uniformly heated. To 
obtain a satisfactory result, the furnace must re- 



ROASTING OF OEES. 97 

volve very slowly. It makes one revolution in two 
to five minutes. The ore is carried up by the re- 
volving furnace to a certain height, whence it falls 
through the flame. The draught through the cylin- 
der carries a part of the finest ore out through the 
flue ; it is therefore necessary to build dust-cham- 
bers between the flue and the chimney. On the 
long side of the cylinder is an opening, closed by 
a lined iron door, through which the ore is dis- 
charged when finished. Through this same door 
the ore is charged. This kind of furnace has been 
in use at the La Dura works for several years, 
where it has given satisfaction. It requires less 
fuel and less labor than the single reverberatory, 
and only of late has it been replaced by a long fur- 
nace. Some are still in use in Colorado Territory. 

57. O'Hara's Mechanical Chain Furnace. Of all 
furnaces, the object of which is a continual dis- 
charge of roasted ore, taken directly from the 
stamp^ without the intervention of manual labor, 
O'Hara's was the only one crowned with practical 
success. Stetefeldt's furnace of late is arranged in 
the same way that is, concerning the direct feed- 
ing from the stamps by machinery. O'Hara's fur- 
nace is in use on Carson River, Nevada. Three of 
these were in operation at Flint, Idaho Territory. 
At present, however, they have stopped for want 
of ore. The construction of O'Hara's furnace is 
shown by an outline drawing, as represented in 
5 



98 



ROASTING OF ORES. 



Fig. 8. The hearth, A, is 104 feet long and nearly 
five feet wide. Eighty feet of this hearth are 
crossed by an arch, B, twelve inches 
high, and connected with three fire- 
places, two, c and d, on one side, and 
one between c and d on the other. 
a is the feeding hearth, provided with 
ore continuously from the batteries. 
The motion of the ore is effected 
by an endless chain, g, passing over 
two chain wheels, one at each end. 
To this chain two oblong flat rings, 
h, are attached, each provided with 
eight shovels or plows so arranged 
that while one of the rings shoves the 
ore toward the center line, the other 
pushes it back again toward the sides 
every three or four minutes (or in 
shorter intervals if more ore is 
charged). The ore not only changes 
its place to the right and left, but it 
also moves forward by degrees, so 
that in the course of six hours from 
the beginning, it commences to be 
discharged at f, passing eighteen 
"^ feet over the cooling hearth, e. On 
both ends of the furnace are iron 
doors hung on hinges, which are 
opened by the rings. After several 
months of operation the hearth or 
bottom appeared in good condition. 



co 




ROASTING OF ORES. 99 

The five batteries, five stamps each, have on 
both long sides endless screws, by which the 
crushed ore is forwarded, in proportion as it is dis- 
charged, to an elevating apparatus. Being lifted 
about fifteen feet, it is conveyed again by endless 
screws along the feeding hearths of all three fur- 
naces, a', and regularly divided and discharged on 
the feeding hearth, a. The ore,- mixed with 100 
pounds of salt to each ton, is spread on iron plates 
before the batteries, (heated by the hot air from the 
furnaces, conveyed through the flue and under the 
plates.) When charged into the battery the ore is 
not further handled till it comes out of the furnaces 
perfectly roasted ( 32). 

There is only one obstacle connected with this 
and other mechanical furnaces. The shoes or 
shovels, touching the sides of the furnace, wear off 
by degrees, leaving a space which is taken up by 
the ore. This part of the ore along the wall hard-, 
ens and increases in amount in the furnace till new 
shoes are put in. By these the crust of one-half 
to three-quarters of an inch thick is broken off and 
carried out. From the Rising Star ore these crusts 
contain nearly just as much chloride of silver as the 
well roasted ore ; they are, nevertheless, disagree- 
able, but some means might be devised by which 
this inconvenience could be avoided. 

58. Stetefeldt's Coasting Furnace. This furnace, 
now being built at Austin, Nevada, is represented 
in Fig. 9, showing a vertical cross section. The 



100 



ROASTING OF ORES. 



furnace at Reno, Nevada, has a dust-chamber in 
place of the flue, b, of Fig. 9, the omission of 




which simplifies the construction without injury to 
the good results of roasting. The furnace has three 
important departments. 1st. The roasting shaft, 



ROASTING OF OEES. 101 

a, twenty-five feet high and five feet wide at the 
bottom, narrowing somewhat toward the top, to 
prevent the adherence of dust to the wall. It is a 
simple shaft of common bricks, built as smooth as 
possible. On the top of the shaft, at a', is placed 
an iron feeder, through which a permanent and 
uniform feeding of the pulverized ore, already 
mixed with salt, is effected. The ore falls on the 
bottom, e, and when half a ton or a ton is accumu- 
lated, it is drawn out through the door, f. 2d. 
The fire-places. There are three gas generators, 
constructed similarly to that of the copper-refining 
furnace at Mansfield, Prussia. The cover is taken 
off and the charcoal introduced. The cover is 
placed again on its frame, which contains sand in 
a groove in order to shut off the draft entirely. 
The slide door near g is drawn out, and the char- 
coal falls on the grate, h, through which as much 
air is admitted as is necessary. There are also two 
canals on each side of the grate, oi\e of which is 
shown by dotted lines, i, both communicating at 
k. Through these canals is regulated the admis- 
sion of the air for oxidizing or burning the car- 
bonic acid, created above the grate, h. In the flue, 
d, air and gas meet together, and the burning pro- 
duct heats the furnace. Two of these generators 
heat the shaft, a ; the mouth of one is shown in 
the drawing by c, the other is on the opposite side, 
and therefore not visible in the plan. The two 
generators are constructed exactly like g, with the 
exception that the flue, d, is not inclined, but hori- 



102 BOASTING OF ORES. 

zontal. The flue, d, as well as the generators above 
the grates, are lined with fire-bricks. 3d. The 
dust-chambers. With the draft, the gases from 
the shaft, with a part of the fine ore dust, pass 
through the vertical flue, b, then through the hori- 
zontal one, m, into a series of chambers, ?i, of dif- 
ferent sizes. The first four chambers, n, are 
smaller than the four following, which are not rep- 
resented in the diagram ; from the last chamber 
the gases draw into the chimney. The dust can be 
removed from the bottom of the chambers through 
the doors, o, o. Almost all the dust is regained, 
and not in a raw condition, as from dust-chambers 
of reverberatory furnaces, requiring re-roasting, 
but perfectly chloridized, which is principally due 
to the auxiliary generator, g y and the longer con- 
tact with the chlorine gases. 

Chimneys and Flues. 

59. The draft in a furnace depends on the 
height of the chimney. The flue or canal between- 
the hearth and chimney has a great influence on 
the draft, as a great deal of heat is taken up by the 
walls, and the draft in the chimney depends on the 
temperature therein up to a certain degree. It 
follows that the longer the flues are, the higher the 
chimney should be. Flues underground, once 
heated, absorb less heat than if exposed to the air. 

Single roasting furnaces, each having its own 
chimney, dispense entirely with long flues. It is 



BOASTING OF ORES. 103 

therefore sufficient to build the chimney twenty to 
twenty-five feet above the level of the hearth, and 
fifteen to eighteen inches square in the clear. 

Underground flues are suitable where many 
roasting furnaces are connected with the chimney. 
They are often built directly under the furnaces, 
two feet wide by three to four feet high. In this 
position the connections between, the main flue and 
those of each furnace are the shortest. Although 
this mode is preferable to flues on the side of the 
furnaces, it is not practicable where deep ash-pits 
are in use. Deep ash-pits are "favorable, partly for 
the reason that it is not necessary to carry out the 
ashes every day or two, but principally on account 
of the grates, which last longer, being better 
cooled by the air. 

The flues are sometimes needlessly carried too 
far out, especially when it is intended to reach 
ascending ground on which the flue continues, re- 
placing the chimney entirely or in part. The ascent 
must be steep, otherwise if the length is in no pro- 
portion to the perpendicular height gained by it, 
taking also the distance from the furnaces to the 
ascending ground in consideration, the absorption 
of heat might neutralize the advantage of the 
ascending flue. On the other hand, the tempera- 
ture in the chimney should not exceed 300 C. = 
572 F. 

A chimney fifty to seventy feet high, and from 
three to four feet square inside, is generally con- 
sidered sufficient for a number of roasting fur- 



104 EXTRACTION OF SILVER. 

naces. It is built of common bricks, sometimes of 
stone. In the latter case the stone must be exam- 
ined to ascertain whether it will stand the heat 
without bursting. The stone work is often lined 
inside with a layer of bricks. 

According to experience, the most advantageous 
proportion between the area of the grate-openings 
and the section of a chimney, is between one to 
one and two to one. The round section is the 
most proper, as there is the least friction with it. 
The section ought to be the same at all heights. 



III. EXTRACTION OF SILVER 

BY WAY OF LIXIVIATION. 

Solving Process. 

60. Under ' ' Solving Process " is to be under- 
stood here, simply roasting with salt, and extracting 
the silver with hyposulphite of lime or of soda, 
without reference to particulars of roasting in 
the Patera or Kiss processes. 

The solving process comprehends, generally con- 
sidered, different modes of extraction, all of which 



EXTRACTION OF SILVER. 105 

are based on the property of the solubility of the 
chloride and sulphate of silver. The extraction of 
the chloride of silver by alkaline hyposulphites was 
proposed by Percy. Patera was the first who made 
use of the hyposulphite of soda for extraction of 
silver in a practical way; his success, however, de- 
pends principally on his modified and complicated 
roasting ( 33). By lixiviation the silver is ex- 
tracted in the Patera, Kiss, Roszner-Patera, Zier- 
vogel, Augustin, and Kustel & Hofman processes. 
The extraction of silver by the solving process is 
simple. The ore is first roasted with salt in the 
usual way, whereby the formation of base metal 
chlorides cannot be avoided entirely. After roast- 
ing, the ore is first subjected to leaching with 
water, in order to extract the base metal chlorides, 
and then with hyposulphite of lime, to extract the 
silver. 

The Extraction of Silver. 

61. After a chloridizing roasting the ore should 
be examined to ascertain the amount of chloride of 
silver contained in it, according to 21. In case 
the extraction should not be satisfactoiy, it is then 
easier to find what the cause is. The ore is then 
prepared for leaching. 

A. First Leaching. The roasted ore contains 

chloride of silver, which does not dissolve in water, 

but generally there are also base chlorides in it, as 

the chlorides of copper, zinc, lead, iron, antimony, 

5* 



106 



EXTRACTION OF SILVER. 



etc., which are soluble. It is the purpose of the 
first leaching to extract these base metals by means 
of hot water. For this purpose the ore is intro- 
duced into a tub or square box of pine wood, the 
planks being one and one-half to two inches thick. 
Fig. 10. 




The boxes must be made as water-tight as pos- 
sible and provided with a filter at the bottom. The 
filter is prepared in two ways, either as represented 
in Fig. 10 by fixing a false bottom, a, provided 

Fig. 11. 




with numerous holes, one-half inch in diameter, 
about one inch above the bottom, b, or as 
Fig. 11 shows, without a false bottom. On the 
bottom, a, is thrown clean rock, quartz or poor ore 
of about the size of a hen's egg, three or four inches 



EXTKACTION OF SILVER. 107 

high ; on this smaller stuff, and finally sand, free 
from mud. In Fig. 11, rock of about the same 
size is thrown directly on the bottom, c, spread 
four inches high, then a few buckets full of rock, 
not smaller than hazel nuts, and only so much of 
it taken as to equalize the surface. This is then 
covered with a piece of canvas and is ready for use. 
The boxes, according to their size and the weight 
of the ore, may contain from one to five tons. 
Generally the ore must not be over fourteen or six- 
teen inches deep, but some ore allows a good 
leaching with twenty-five inches. 

The roasted ore, generally without sifting, is 
charged into the boxes, and the surface spread 
evenly, leaving about six inches space from the top 
for the reception of the leaching water. The hot- 
ter the water is, the sooner it dissolves the soluble 
salts and the quicker the leaching progresses. It 
is conveyed through the pipe, d, and falls on a 
piece of canvas, whence it spreads equally and 
gently over the ore. The water soon reaches the 
bottom and begins to flow out through the pipe, e, 
into the trough, f. 

In the beginning, the leaching water at e is 
highly charged with base metal salts, and shows a 
green color if there is much copper in the ore. 
The water is kept running in a continual stream 
till it reaches nearly to the rim of the box, when 
the influx and the efflux are equalized. After one 
or two hours a glass full of the liquid, at the pipe, 
e, is taken, and a few drops of sulphide of calcium 



108 EXTRACTION OF SILVER. 

(or of sodium) added. If a precipitate falls, of a 
dark or light color, the leaching must continue ; 
but it is not necessary to continue until no precipi- 
tate at all is perceived, as it requires some time 
perhaps an hour before all the water runs out 
after the pipe, d, is closed. The water which 
comes out last must be free from salts. This first 
leaching takes from two to four hours, sometimes 
longer. 

B. Second Leaching. As soon a*s the ore is 
freed from the base chlorides soluble in water, a 
solution of hyposulphite of lime ( 70) is led in 
from a tub or tank, on the ore, in order to dissolve 
the chloride of silver. This leaching is conducted 
like the former. It depends on the amount of 
silver how long this work continues from eight to 
twenty hours. The clear cold solution, containing 
the chloride of silver in the form of a double salt, 
has a very sweet taste, and is conveyed through a 
trough or india rubber hose into a precipitating 
tub. Very rich ore, containing 12 to 15 per cent, 
of silver, would require forty-eight hours leaching, 
and even then it would be necessary to subject the 
ore to a second leaching with the hyposulphite, 
with an intermediate roasting with green vitriol 
and salt ; for, with the best work, if 95 per cent, 
are extracted, the tailings would still appear suffi- 
ciently rich for this, containing about 200 ounces 
of silver per ton. Ores containing $350 per ton 
are often leached out perfectly in twelve hours. 
The end of the lixiviation is ascertained in the 



EXTRACTION OF SILVER. 109 

same way as in leaching with water, using the sul- 
phide of calcium. If no precipitate is obtained 
the extraction is finished. 

The color of the precipitate is a black-brown. 
The presence of base metals changes the color 
somewhat. Iron makes it black ; copper, red- 
brown ; lead and antimony, light red-brown, etc. 
The silver is first dissolved, especially if a diluted 
solution of hyposulphite of lime is used ; and for 
this reason the first precipitate is the richest in 
silver. Ore containing a great deal of lead 
especially if the roasting was so conducted that a 
large part of it remained as sulphate of lead, which 
is not soluble in the leaching water will give in 
the beginning of the leaching with the solvent a 
precipitate of silver with some lead ; afterwards, 
however, the silver diminishes, so that the precipi- 
tate of lead finally appears free of silver. Besides 
the sulphate of lead, sub-chlorides and oxy-chlo- 
rides are formed during the roasting which are not 
soluble in water, but are dissolved by the hyposul- 
phite of lime ; for this reason always some base 
metals will be found in the precipitate. 

In case rebellious ores are treated, and hot water 
is used for the extraction of base chlorides, a bet- 
ter silver is obtained if the ore is cooled down by 
cold Water before the cold and diluted solvent is 
applied. Purer ores may be treated with a warm 
solution of the solvent. 

When examining the tailings as to the amount 
of silver left therein, it must be remembered that, 



110 EXTRACTION OF SILVER. 

after leaching out a quantity of metals by water 
and the solvent, the ore lost a considerable part of 
its original weight, and that consequently one-half 
ounce of such tailings taken into assay will always 
give a larger silver button than there ought to be. 
A true assay of leached tailings is made if half an 
ounce of the same ore is leached on a filter with 
hot water and hyposulphite of lime, in the same 
way as the ore on a large scale, washed with water, 
dried and weighed. The weight found after leach- 
ing must be taken for half an ounce in assaying 
the tailings. 

The residue, or tailings in the leaching box, must 
be removed now as valueless. The sides of the 
leaching boxes are from eighteen to twenty-four 
inches above the bottom, and being from six to 
eight feet square in the clear, the removing of the 
tailings by means of chloride is easily effected. 
The men must be careful not to dig too deep, 
otherwise the filter will be injured. It is quite 
proper to fix wooden staves, as long as the box re- 
quires, on top of the filter. These staves are one 
inch wide and one-half of an inch thick, and are 
placed from four to five inches apart, so as to pro- 
tect the canvas or filter against the shovel. In 
Fig. 11 the staves are laid upon the canvas. 

The leaching boxes or tubs may be arranged so 
that, being tipped over, the whole charge falls out 
at once. In this case the filter must be made in a 
different way from that described above. Eocks 
are not serviceable here. On the false bottom, n, 



EXTRACTION OF SILVER. 



Ill 



of Fig. 10, a layer of thin, leafless switches is 
placed, and on this another one crosswise, then 
covered with a piece of canvas, and secured with 
some staves to prevent the falling out of the whole 
filter when turned over. 



Precipitation of the Silver. 

62. The liquid of the second leaching is con- 
veyed through a trough or india rubber hose to the 

Jfiff. 12. 




precipitating tanks, of which three or four are em- 
ployed. If tubs are used, which for this purpose 
are the best, they are from three to four feet in 
diameter, and four feet high. The tanks or 



OF THB 

TJNIVERSIT 



: EXTRACTION OF SILVER. 

boxes have a rectangular shape of about the same 
capacity, the bottom being inclined toward the 
middle, as shown by Fig. 12. The hyposulphite 
of lime, as it comes from the leaching tanks, is 
conducted into these until they are more than two- 
thirds full. . The trough or hose is then changed 
to discharge the liquid into the next precipitating 
tub, while the precipitation of the first commences. 
The liquid used for precipitating the silver is 
sulphide of calcium ( 69). It is poured in until 
all the silver is supposed to be precipitated, and at 
the same time the solution is stirred vigorously. 
Treating always the same kind of ore, the required 
quantity of the precipitating agent is soon learned. 
The black precipitate sinks to the bottom, and the 
workman now dips a little of the clear liquid 'out 
in a glass tube or tumbler, and adds a few drops of 
the sulphide of lime. If a dark precipitate or a 
dark color is produced, it shows that there is still 
silver in the liquid, and more of the agent must be 
added ; but if on the contrary no precipitate is ob- 
served, there is either enough or too much of the 
sulphide. To prove this, some of the silver-hold- 
ing liquid is added to a test, taken from the tank 
under treatment. If in this case a precipitate is 
formed, silver-holding liquid must be carefully 
added to the tank until no reaction is produced. 
This work, delicate as it seems, is easily learned by 
the workmen. If a little silver should be left in 
the liquid, it is not injurious, neither is the silver 
to be considered as lost, because the same liquid is 



EXTRACTION OF SILVER. ' 113 

used over again ; but a small excess of the sulphide 
of calcium would cause a loss in silver, as it pre- 
cipitates sulphide of silver in the leaching tank in 
the mass of ore, which is not dissolved again. The 
precipitation is performed in a short time, requir- 
ing about fifteen minutes for each tank. The stir- 
ring must be executed with vigor. Wooden grates 
fixed to a vertical stem will answer the purpose. 

The clear solution above the settled precipitate 
is pumped or elevated to the reservoir, whence it 
was conveyed on the ore. It is now ready to be 
used again. The sulphide of calcium having per- 
formed its duty in precipitating the silver, is turned 
into hyposulphite of lime, thus replacing all of the 
solvent. 

To prevent small floating particles of silver from 
being elevated with the liquid, it is well to allow 
sufficient time for the precipitated silver to settle. 
For this reason it is better to have more precipi- 
tating tanks or tubs. It is not necessary to re- 
move the silver after each precipitation. The clear 
liquid can be drawn off, by means of a syphon, 
from all the precipitating tubs into a general re- 
ceiver, whence it may be .pumped up. After the 
solvent has been removed, the precipitated silver 
can be drawn off through the pipe, d, Fig. 12, di- 
rectly into canvas bags. 

Treatment of the Precipitated Silver. 

63. The black precipitate of sulphide of silver 
is conveyed directly into filters made of canvas, 



114 EXTRACTION OF SILVER. 

either in the shape of pointed bags, like those used 
for amalgam, or in the shap'e of common bags. As 
soon as all the liquid runs out, pure water (if pos- 
sible, warm) is poured on the silver, and this re- 
peated several times till no taste is observed in the 
filtering water. The precipitate, while still in the 
bags, is placed beneatn a screw press and the fluid 
pressed out as completely as possible. The black 
silver cakes are then taken out and dried in a warm 
room or in a drying oven. For the purpose of 
burning off the sulphur, the dried sulphide is in- 
troduced into a muffle or other calcining furnace, 
and heated till the sulphur commences to burn 
with its known blue flame. When this disappears 
the heating must continue at a dark red heat for 
one or two hours. By this operation the cakes are 
reduced almost entirely to metallic silver, generally 
covered with threads of silver ; sometimes an in- 
tense green color is assumed by pieces remaining 
in the furnaces over night. 

The burned cakes are now prepared for smelting 
in crucibles. They are placed in black lead cru- 
cibles, according to the size, up to three hundred 
pounds, and fused. All the sulphur was . not 
driven out by the preceding operation. The re- 
maining part must be removed by placing metallic 
iron ( 5 ; d) in the fused metal ; thereby iron matt 
is formed, which rises to the surface and is 
skimmed off. The surface of the silver is then 
cleaned by adding some bone ash and borax, or 
borax alone, which is also skimmed off and the 



EXTRACTION OF SILVER. 115 

silver dipped out or poured out into moulds. Ac- 
cording to the careful treatment in the roasting 
process, and the nature of the ore, the silver will 
be from 800 to 950 fine. 

Mr. O. Hofmann, in need of sulphur for the 
production of sulphide of calcium, used to calcine 
the dried sulphide of silver in iron retorts. In this 
way he obtained a large proportion of sulphur as a 
fine sublimate. This could be done also in a proper 
muffle furnace, so arranged that after all obtaina- 
ble sulphur had sublimated in a receiver this could 
be removed and the calcination continued under 
access of air. 

Precipitation of Copper contained in the 
Ore and of a small amount of Silver 
leached out with the Copper. 

64. Having refractory ore under treatment, it 
is generally the case that copper is also found in it. 
While roasting, the presence of copper is favorable 
for the chlorination of the silver, but copper ores 
require some more salt, especially if it is intended 
to save the copper also. The more chloride of 
copper formed, the more will be found in the so- 
lution while leaching it with hot water. In order 
to convert all the copper into a chloride, it would 
take at least one and a half pounds of salt to each 
pound of copper ; and considering other base 
metals, lime, etc., all of which absorb chlorine, 
while a considerable part escapes useless, the above 



116 EXTRACTION OF SILVER. 

quantity has to be doubled. For this reason no 
special attention can be paid to the copper ; only 
that part of it can be extracted which is converted 
into a chloride during roasting under the usual 
circumstances. The chloride of copper transfers a 
part of its chlorine to the silver and other metals 
( 23), and is reduced thereby to a sub-chloride ; 
if there is sufficient salt in the furnace it is raised 
again to a chloride. This sub-chloride (Cu 2 Cl) is 
not soluble in water ; it remains in the ore during 
leaching. 

65. The different chlorides, being removed in 
the first leaching ( 61), are principally those of 
copper, iron, lead, antimony and zinc, besides some 
undecomposed salt. The first quantity of hot 
water introduced into the leaching box is, of 
course, most saturated with the named salts, and 
they have the property of dissolving, also, some 
chloride of silver. The dissolved silver precipi- 
tates again as soon as it becomes diluted with more 
water. There is, therefore, no diniculty in regain- 
ing the silver which is thus leached out. The 
amount of silver carried out by the leaching water 
varies from 0.5 to three per cent. Not only the 
chloride of silver, but also those of lead and anti- 
mony, are precipitated by dilution with water. 
There are two ways of regaining this silver. 

Mr. O. Hofmann adopted an ingenious plan for 
this purpose, by conveying the hot water, under a 
slight pressure from below, through the pipe, e, 



EXTK ACTION OF SILVEE. 117 

( 61, Fig. 10), by attaching to it a rubber hose. 
The water rises through the ore from e up to d, and 
as soon as it reaches within two inches of the brim 
of the box, the hose is removed from e, and the 
water admitted through d. The concentrated so- 
lution, containing dissolved silver, is now above 
the ore, and being diluted with water from d, lets 
the chloride of silver fall as a precipitate on and 
throughout the ore. 

The other plan is the precipitation of the silver, 
together with the chlorides of lead and antimony, 
outside of the leaching box. This mode is prefer- 
able to the former when a great deal of lead and 
antimony is in the ore ; for if precipitated in the 
box, a great part of it will be dissolved by the hy- 
posulphite of lime and then precipitated as sul- 
phides with the silver, making this impure and con- 
suming much of the precipitating agent. As soon 
as the chlorides flow into the trough, f, below e, 
into which several leaching boxes discharge their 
fluids in different degrees of dilution, the gradual 
precipitation commences. The precipitate is white 
and adheres to the trough through the whole 
length of it. These chlorides are the richest, and 
contained, at Flint, Idaho, 9 per cent, of silver; the 
balance was principally lead and antimony. The 
precipitate deposits on all bodies offering a surface. 
For this purpose a box must be constructed, as 
represented in Fig. 13, which shows the top view 
or ground plan. The sides, a, of a wooden box, 
six feet by six, or ten by six, are six inches high, 



118 



EXTRACTION OF SILVER. 



the front, b', and partitions, b, four inches high, 
leaving a space of six inches between them. These 
spaces are filled with shavings representing an im- 
mense amount of surface for the chlorides to de- 
Fig. 13. 




I 
3k 

,i 

^ 



j 



posit on. The fluid entering the trough, c, contains 
now a purified copper solution. Chloride of iron 
is also with the copper in solution, but does not 
prevent the copper from precipitating. 

66. The white precipitate, when accumulated, 
is taken out, placed in filtering bags, with or with- 
out the shavings, and washed with clear cold water, 
in order to get rid of the copper solution. The 
silver can be extracted in two ways : The simplest 
mode is the application of hyposulphite of lime. 



EXTRACTION OF SILVER. 119 

The sediment is taken out from the filtering bags 
and charged, while wet, into a filtering box of a 
proper size, arranged like Fig. 11, 61. The 
hyposulphite of lime, in a cold condition, is poured 
over it and managed as with the ore with the 
second leaching, 61. The silver-holding fluid 
may be conveyed into the precipitating box, Fig. 
12, 62, and treated with the solution from the 
ore. The liquid from the bags is examined from 
time to time with sulphide of calcium. In the be- 
ginning the precipitate appears dark, being mostly 
silver ; but when it is perceived that the precipi- 
tate assumes a light yellow color, too much of lead, 
zinc and antimony is being carried out, and the 
leaching must be stopped. The residue in the fil- 
ter box contains still some silver. 

The other mode of extraction is more perfect, 
but also more expensive and more troublesome. 
After the copper has been washed off, the contents 
of the bags are taken out and dried. It is then in- 
troduced into large crucibles and smelted with an 
addition of soda-ash. The reduced metal, if some 
lead occurs in the ore, must be separated by means 
of cupellation, resulting in clean silver and litharge. 

67. The chloride of copper running from the 
box, Fig. 13, is led into a reservoir in which old 
iron is suspended. The copper precipitates in a 
metallic state on the iron, and about eighty-eight 
parts of iron go into the solution in place of one 
hundred parts of copper ; consequently, as each 



120 EXTRACTION OF SILVER. 

one hundred pounds of pure copper require eighty- 
eight pounds of iron, the calculation as to the 
necessary amount of iron could be made easily if it 
were not for some other chlorides which may still 
be in solution, and which also require iron for pre- 
cipitation. Wrought iron is preferable to cast 
iron, and gray cast iron is better than white; but 
all these sorts precipitate the copper, and it de- 
pends to a great extent on the price as to what 
kind of iron is chosen. 

The most effective precipitant is the iron-sponge 
or finely divided iron, obtained by heating pulver- 
ized iron ore or roasted iron pyrites with charcoal 
powder in a proper reverberatory furnace, or in 
iron pipes or cylinders without admitting air. By 
these means the iron oxide is reduced to metallic 
iron, which precipitates the copper in a few 
minutes. Using old iron, the precipitation will be 
effected much more quickly than in tanks if an ar- 
rangement is made like Fig 13, putting the old 
iron in the place of the shavings. This would be a 
continuation of Fig. 13, but the box must be three 
or four times as long before it reaches the tank or 
reservoir. 

To find out whether there is yet copper in the 
liquid, the best test is to take some drops on a 
piece of platinum, and to place a small, clean piece 
of zinc on it. The copper immediately appears of 
a bright red color. But for practical use, a clean 
piece of iron dipped into the liquid will also show 
a red coating if there is enough copper in it to 



EXTRACTION OF SILVER. 121 

make it remunerative to continue the precipitation. 
The water contains now principally chloride of 
iron, and is discharged. If by some cheap means 
the water could be evaporated, the remaining chlo- 
ride of iron could be used in roasting ores without 
salt. 

Where old iron commands a high price, the cop- 
per can be precipitated with a brine of ashes or of 
lime ; but in this case the iron also falls with the 
copper. The brine for this reason cannot be ad- 
vantageously adopted where a great deal of iron is 
in the ore, or the roasting must be directed so as 
to decompose the chloride of iron. 

Quality of Ores fit for the Solving Process. 

68. There is no process so suitable for all kinds 
of ores as the solving process. G-enerally consid- 
ered, all silver ores can be treated by the solving 
process which are subjected to the pan amalga- 
mation after roasting ; but in many instances 
especially with the rebellious ores a better result 
is obtained by this than by working in pans. The 
great advantage of this process is cheapness. 
Boasting of course is indispensable except with 
chloride ores ; but neither pans and the required 
power, nor quicksilver, are used, and for this reason 
less capital is necessary to put up reduction works. 
All the cupreous silver ores of Cerro Gordo, Yellow 
Pine, Montgomery, and of the other new silver 
districts, can be treated to great advantage by the 
solving process. 
6 



122 



EXTRACTION OF SILVER. 



rr 

1 _\ 





Two objections have to be considered. First, 
there is more water required than for pan amalga- 
mation at least this is the case with rebellious 

ores ; but the quantity 
of water depends on the 
quantity of base metals 
in the ore, and also on 
the arrangement of the 
leaching boxes. One 
box, containing one ton 
of ore, requiring three 
hours leaching, may con- 
sume 250 gallons of 
water ; three boxes of 
the same size would 
take three times as much 
water if placed on the 
same level ; but by ar- 
ranging the boxes in a 
less favorable position, 
one above the other, as 
shown in Fig. 14, only 
one-half of the quantity 
of water is needed. It 
takes more time to leach 
three boxes together than 
a single one. Leach- 
ing the ore with hypo- 
sulphite of lime, the supply of this in the first box 
must be stopped if no more silver comes out, and the 
solution carried, by means of hose, to the second, 



EXTRACTION OF SILVER. 123 

and then to the third box, if it should be neces- 
sary; so that while the second box is yet under 
leaching, the first can be discharged and a new 
charge introduced. All three boxes should receive 
clean water at the start. This arrangement should 
be adopted only when rendered necessary by the 
scarcity of water. To have all leaching boxes on 
a level is preferable. 

The other objection is confined to a certain class 
of ores containing clay and lime. If pulverized, 
so much fine pulp will be produced that the leach- 
ing is impossible. It is not advisable to crush the 
ore coarser than will allow of its passing through a 
sieve of forty holes to the inch, in some cases, 
perhaps, through thirty-five holes ; and if with 
such crushing a fine clay pulp is produced, the ore 
is unfit for all leaching processes, unless wet crush- 
ing is adopted, in order to separate the slime from 
the sand, as Mr. O. Hofmann was compelled to 
arrange in Trinidad, Sonora. In this case, a sepa- 
rate drying for the purpose of roasting is not 
necessary if long furnaces are in use. It is not 
unlikely that for similar ore and pan tailings an 
agitating filtering box could be constructed which 
would render the leaching possible. 

Sulphide of Calcium. 

69. Sulphide of calcium for the precipitation 
of silver is preferable to the sulphide of sodium, 
principally for the reason that its manufacture is 



124 EXTRACTION OF SILVER. 

cheaper and more eas} 7 3 but also on account of the 
quality of the precipitated silver, which is easier to 
wash, to press and to desulphurize. The sulphide 
of calcium is easily obtained and manufactured on 
the ground where the mill is situated. The articles 
required for this purpose are brimstone (worth about 
four cents per pound) and burned lime. The sul- 
phide is formed only from caustic lime, consequently 
more is obtained from fresh burned lime. Of this 
a certain quantity is charged into an iron kettle, 
water added, and then the pulverized sulphur. The 
proportion of sulphur and lime depends on the 
quality of the latter. The purest quality of lime 
from Santa Cruz, Cal., for instance, takes one 
pound of sulphur to 1.33 of lime. Of poorer quali- 
ties of lime it is better to take three pounds to one 
of sulphur and about ten parts of water. It is 
kept boiling for two or three hours, stirred with 
woo.den shovels, and then discharged into a filter- 
ing box, prepared like Fig. 10, 61. The clear, 
dark, yellow-red sulphide of calcium comes out 
from below the filter, and can be kept in iron ves- 
sels. The liquid ought to be between 5 and 6 
Beaum4. The residue is washed with water, 
whereby a diluted fluid is obtained, which is used 
with the lime of the next charge. Mr. E. Smyth, 
in La Dura, Mexico, treats the lime and sulphur 
with steam. This has the advantage of dispensing 
with the stirring, and may be performed also in 
wooden vessels. The steam replaces the fire and 
has no chemical influence on the quality of the sul- 



EXTRACTION OF SILVER. 125 

pliide ; the precipitating capacity of the latter with 
reference to the volume depends only on the con- 
centration which is expressed by the degrees of 
Beaume's hydrometer. 

One pound of lime (Santa Cruz quality) gives 
sufficient sulphide of calcium to precipitate one 
and a half pounds of silver. 

Hyposulphite of Lime. 

70. The hyposulphite of lime as a solvent of 
chloride of silver has a great advantage over a hot 
solution of salt. It can be applied diluted and 
cold, and dissolves a great deal more of the chlo- 
ride than does the salt, of which nearly sixty-eight 
pounds are required to dissolve one pound of chlo- 
ride of silver, while only two pounds of the hypo- 
sulphite are needed to dissolve the same amount of 
the chloride. 

The hyposulphite of lime is produced by convey- 
ing sulphurous acid into sulphide of calcium till 
it appears entirely colorless. It is also formed if 
from a concentrated brine, obtained from lixivi- 
ating roasted ore with hot water, all chlorides are 
precipitated by sulphide of calcium. After the 
precipitated sulphides have settled, the clear fluid 
can be used to dissolve the chloride of silver. The 
simplest way, however, is to buy hyposulphite of 
soda, and to commence the second leaching ( 61) 
therewith, precipitating with sulphide of calcium. 



126 EXTRACTION OF SILVER. 



Patera Process. 

71. The most delicate operation in Patera's 
-process is the preparation by roasting, as described 
in 33. The chloride of silver formed during the 
roasting is dissolved by a cold solution of hyposul- 
phite of soda, after all soluble base metals have 
been first leached out with hot water ( 61). 'Two 
parts of the hyposulphite of soda dissolve one part 
of chloride of silver, forming a soluble double salt. 
The tubs in which the ore is lixiviated with the hy- 
posulphite of soda are small, receiving only 200 
pounds of roasted ore. The extraction of silver is 
performed in the same way as described under 61 . 

Kiss Process. 

72. This process extracts silver and gold to- 
gether. Boasting the ore, as explained in 38, 
the gold is transformed into such a state as to ren- 
der it insoluble in water. After roasting, the ore 
is placed in filtering tubs and washed with water 
to remove the base metals. After this a solution 
of hyposulphite of lime is conveyed on the ore, by 
which the gold and silver chlorides are dissolved 
and carried off into precipitating tubs. As soon 
as the sulphide of calcium is introduced, the gold 
and silver are precipitated as sulphides. The pre- 
cipitation of both metals in a metallic condition is 
not admissible, for the reason that the hyposul- 



EXTRACTION OF SILVER. 127 

phite of lime is decomposed if metallic copper is 
employed for precipitation. The results of 'Kiss's 
methods, practiced in Hungary, were not altogether 
satisfactory. 

Patera and Rceszner Processes. 

73. The object of these processes is, like that of 
the preceding, the extraction of silver and gold 
together. The ore is first subjected to a chloridiz- 
ing roasting, by which the silver is converted into 
a chloride, while the gold remains mostly in me- 
tallic condition. The leaching liquid is prepared 
by conveying chlorine gas through a cold concen- 
trated solution of salt to saturation. This chlo- 
ridized solution dissolves silver, gold and copper 
at the same time. The roasted ore is charged into 
tubs with false bottoms, and the cold solution of 
salt and chlorine introduced. Silver ores treated 
after this method in Hungary gave 98.94 per cent, 
of the silver, all the copper and nearly all the 
gold. An experiment on five tons of ore gave a 
clear profit of seventy-five florins, compared with 
the amalgamation. 

Roszner roasts the ore with salt, extracts a part 
of the silver by Augustin's method with a hot so- 
lution of salt, and treats the residue alternately 
with a solution of salt and chlorine, and with a hot 
concentrated salt solution for the extraction of gold 
and the remainder of the silver. 



128 EXTRACTION OF SILVER. 



Kustel & Hofmann Process. 

74. Auriferous silver ores are roasted, as de- 
scribed under 35. They are then subjected to a 
process differing from the preceding ones in ob- 
taining separately the copper, gold and silver. 
After roasting, the ore must be moistened on the 
floor, by conveying water through a spout, and 
mixed with shovels, so as to get it uniformly moist, 
but not so wet as to interfere with sifting, which, 
however, is not always necessary. The ore is then 
put into tubs or boxes, as represented by Fig. 10 
or 11, 61, but provided with covers, which can 
be easily screwed air-tight on the rim of the box, 
having india rubber between the box and the 
covers. The ore should not reach the brim of the 
box, but leave a space of four or five inches at 
least above the ore, as a chamber for surplus chlo- 
rine. When the box has been filled, the cover is 
screwed or otherwise fastened on the box, and the 
chlorine gas admitted ( 11). 

The chlorine gas is generated in a leaden gen- 
erator of the construction shown in Fig. 15, which 
represents a cross section. It is a circular tub, 
with an outer ring, a, six inches deep, for the re- 
ception of the cover, b. There is a similar ring, c, 
on the top of the cover, which receives the collar 
fastened to the leaden stirrer, d. There is also a 
short leaden pipe, e, bent in the shape of an s, 
through which the sulphuric acid is introduced. 



EXTRACTION OF SILVER. 



129 




Another lead pipe, /, conveys the chlorine out of 
the generator. The vessel is uncovered, and for a 

charge of three tons of 
roasted ore the follow- 
ing ingredients are 
introduced : Thirty 
pounds of peroxide of 
manganese (pulver- 
ized); thirty to forty 
pounds of common 
salt, according to qual- 
ity ; seventy-five pounds 
of sulphuric acid, of 
sixty-six degrees ; and 
forty-five pounds of 
water. Salt, manga- 
nese and water are in- 
troduced first, and the 
generator covered. The 
two rings, a and c, are 
filled with water, in 
order to close the gen- 
erator air tight. The 
sulphuric acid is now 
charged through the 
funnel, e, in small amounts ; twenty to twenty- 
five pounds are sufficient to evolve the chlo- 
rine and the required heat. When the evo- 
lution of chlorine becomes weaker, twenty pounds 
more of acid are administered, and after some time 
the rest of the seventy-five pounds. It will be 
6* 




130 EXTRACTION OF SILVER. 

necessary now to build a fire beneath the gener- 
ator, which is placed on tiles that the heat may 
not injure the leaden bottom, which is made of 
sixteen-pound sheet lead, while the sides are of 
eight-pound. 

The chlorine is not led directly to the ore, but 
through a purifying apparatus, as represented in 
Fig. 16. An ordinary wash basin or a similar ves- 
sel receives the lead pipes (three-quarter inch). 
One of these, b, conveys the chlorine from the gen- 
erator ; the other, c, leads the gas to the ore-box. 
Both ends are covered with a bottle, a, the bottom 
of which is cut off (by meaus of a string). Clean 
water is poured in the basin so as to cover the end 
of the pipe, 6, about one inch high. The -gas is 
forced to pass through the water, by which the 
muriatic acid is taken up. Both lead pipes can be 
provided with rubber hose for connections with 
generator and ore-box. The passing of the chlo- 
rine through the water shows distinctly the rapidity 
of the evolution of the gas in the generator, and 
indicates when there is need of more acid or of the 
application of heat. The water in the basin ab- 
sorbs more chlorine when cold about 2J per cent. 
of its volume before it is saturated. A continual 
stream of water is therefore improper. 

The purified chlorine is now conducted through 
the pipe, e, of Fig. 10 or 11, 61, to which the 
rubber hose or lead pipe from the purifying appa- 
ratus (Fig. 16) is attached. The chlorine goes 
through the whole mass of the ore, driving out the 



EXTRACTION OF SILVEK. 131 

lighter air through a hole in the cover till the gas 
itself comes out at the same hole. A glass rod is 
dipped from time to time into ammonia, and held 
before the hole. As soon as white fumes appear 
on the rod, it is proof that the box is filled with 
chlorine. The hole is now closed and the fire be- 
low the generator removed. In this condition the 
ore remains for twelve or fifteen hours. The whole 
arrangement must be examined with ammonia, to 
see that there is no loss of chlorine. According to 
the amount and quality of the gold, it may be neces- 
sary to allow the chlorine to operate for eighteen 
to twenty hours. 

Lixiviation. It is generally the case, and ought 
to be so, that only a part of the chlorine is con- 
sumed, while the rest is unchanged. For the sake 
of economy and for sanitary considerations, the free 
escape of the surplus gas should not be allowed ; 
but this should be utilized for the same purpose 
that is, for chlorination. This gas is easily trans- 
ferred to another box prepared with moistened ore. 
Mr. O. Hofmann inserted a rubber pipe into the 
hole of the cover of the ore-box, and joined the 
other end with a pipe, e, of another box, as in Fig. 
10. Through the pipe, e, of the already chlori- 
dized ore, he admitted the water, which, entering 
the box, displaced the chlorine and forced it into 
the other vat. The water,^ however, dissolves a 
part of the chlorine, and for this reason it is better 
to convey the surplus gas over by suction, which is 
easily effected by an air-tight tub in which a 



132 EXTRACTION OF SILVER. 

vacuum is created by the discharge of water. This 
way is the more advisable, as a delay, by which the 
leaching water is longer in the box, is injurious to 
the gold. When the gas is removed, hot water is 
introduced, and the leach, containing gold, cop- 
per, etc., led into the precipitating box. An ad- 
dition of dissolved sulphate of iron will precipitate 
the gold in metallic condition. The sulphate of 
iron is either procured as an article of commerce, 
or prepared by throwing old iron into diluted sul- 
phuric acid. 

After several hours (from eight to twelve) the 
clear solution is drawn off from the precipitated 
gold, and if the copper is to be extracted, conveyed 
to other vats and treated as described in 67. The 
liquid, coming out through the filter of Fig. 10, 
must be examined in the beginning with a clear so- 
lution of sulphate of iron, and as long as a dark 
color of precipitated gold is perceived, the leach is 
allowed to run into the gold precipitating tub ; if 
there is no precipitate observed, the leach is di- 
rected into- the copper vats. The liquid is now ex- 
amined with sulphide of sodium or of calcium, and 
proceeded with exactly as shown in 61, " Second 
Leaching." 

Augustin Process. 

75. This process is not in use at present for 
silver ores, but for products of smelting. By this 
method the chloride of silver, which is formed by 



EXTRACTION OF SILVER. 133 

way of roasting ( 30-31), is dissolved in a hot so- 
lution of salt, and precipitated by metallic copper. 
One part of chloride of silver requires sixty-eight 
parts of salt, to be dissolved. 

Extraction of the Silver from Copper Matt and 
Black Copper. The principal aim with these mate- 
rials is the oxidation of the copper as perfectly as 
possible, and then the chlorination of the silver. 
There are wooden leaching tubs of a small size 
two feet eight inches in diameter, and nearly four 
feet high fixed on wheels and arranged in one 
row. Into these tubs, which have false bottoms, 
the roasted stuff is introduced about 800 pounds 
in each. Ores containing different kinds of earths 
cannot be lixiviated at a depth of over three feet; 
the metal oxides, however, allow the water to pass 
freely. This is also the case with roasted, concen- 
trated, or pure sulphurets. Hot solution of salt is 
now allowed to flow through a trough in each tub. 
The salt penetrates the powder, dissolves the chlo- 
ride of silver, and carries it through the filter at 
the bottom of the tubs, flows off to a reservoir, and 
from here, after the particles which may escape 
through the filter have settled, into a series of ves- 
sels one above the other. These are provided 
with double bottoms. The two uppermost rows 
contain cement copper, six inches deep ; the low- 
est, metallic iron. 

The fluid deposits its silver principally in the 
first tub, dissolving at the same time an equivalent 
amount of copper. Some silver which escapes pre- 



134 EXTRACTION OF SILVER. 

cipitation falls with the cupreous fluid into the next 
tub below, where the rest of the silver is taken up 
by the copper. In the third vessel the copper is 
precipitated by the iron. The brine, freed from 
silver and copper, is pumped up into the reservoir, 
heated and used again. The cement copper ob- 
tained in the last tub is placed back in the upper . 
two. The brine circulates in the tubs until a 
bright copper plate is not coated with silver when 
held in the fluid from the leaching tubs. The 
residue, which is mostly copper-oxide, is removed, 
and an .average sample taken and assayed. If it 
should contain over eight ounces per ton, it must 
be roasted over and again lixiviated. 

The precipitated silver is taken out once a week 
and treated with muriatic acid for the purpose of 
dissolving the copper particles which remained 
with the silver. After this it is washed with water 
till all traces of the acid disappear, then pressed 
into balls, dried and melted. 

Ziervogel Process. 

76. Like the preceding, Ziervogel's extraction 
of silver is not applied to silver ores, but only to 
copper matt. The roasting ( 43) is very delicate, 
and it is more difficult to obtain a satisfactory re- 
sult with silver ores than by a chloriclizing roast- 
ing. The silver in this process is converted into a 
sulphate, which is soluble in water, thus dispens- 
ing with the expensive salt brine. The pulverized 



OF THB 

TJNIVERSr 



EXTRACTION OF SILVER. 



and properly roasted copper matt* is charged into 
laaching tubs, 500 pounds in each, and hot water 
admitted. As soon as the water begins to flow out, 
the hot water is made a little acid by admixture of 
some sulphuric acid. The lixiviation continues 
until a sample of the fluid remains clear if a so- 
lution of salt is added. The silver-holding brine 
is conveyed into a large reservoir, thirty feet long, 
where it clears of impurities, which accidentally 
come out of the leaching tubs, and falls from this 
Reservoir through a series of cocks into the precipi- 
tating tubs. On the false bottom is a layer of ce- 
ment copper, and upon this fifteen to twenty cop- 
per bars of 250 pounds weight. Each is fourteen 
inches long, five inches wide and one inch thick. 
The liquid loses most of its silver in these tubs, 
and flows then through a trough fifteen inches 
wide, lined with sheet lead and having a layer of 
copper pieces on the' bottom, into five vats filled 
with copper, where the balance of the silver is de- 
posited. 

The desilverized brine comes now into a reservoir, 
whence it is pumped up into a large leaden pan 
and heated again by means of steam. Above this 
pan is a leaden vessel, out of which about thirty 
drops of somewhat diluted sulphuric acid drop into 
the liquid every minute. The acid prevents the 
separation of basic salts. The silver is taken out 
of the precipitating tubs every day. With it occur 
some copper and gypsum. The larger particles of 
copper are separated by washing, exposed for six 



136 EXTRACTION OF SILVER. 

or seven days to leaching with diluted sulphuric 
acid, and finally washed with hot water. The sil- 
ver is from 860 to 870 fine. After drying, it is re- 
fined in a reverberatory furnace. 

Once a year the brine is brought into contact 
with iron, in order to precipitate the copper. The 
purer part of the cement copper is used for the 
silver precipitation, and the finer part is delivered 
for smelting. 

The Leaching Process.* 

77. Under this name is understood a prepara- 
tion of the ore applicable for the pan amalgamation. 
Its description, therefore, does not belong here 
strictly, but the leaching itself has so close a con- 
nection with the preceding manipulations that this 
part alone may be described without mentioning 
the further treatment by amalgamation. 

It is a known fact that, in treating refractory ore in 
pans by amalgamation, of course by way of roast- 
ing, some very annoying things are encountered, 
and amongst them principally, the great loss of 
quicksilver, amounting sometimes up to ten or 
twelve pounds per ton of ore ; the rapid destruction 
of pans, which compelled many mills to use 
wooden sides fixed to the iron-pan bottom, a meas- 
ure which saves the pans at the expense of quick- 
silver; and the very base bullion which results from 

*The Leaching Process is patented, as an application for pan and 
barrel amalgamation, by G. Kustel. 



EXTRACTION OF SILVER. 137 

such a treatment. In some instances it happens 
that a great deal, sometimes over fifty per cent. , of 
iron goes into the amalgam, rendering the contin- 
uation of the amalgamation impossible. The re- 
sult of the amalgamation of base metals is always 
a certain loss of silver, which would have amalga- 
mated if the base metals were out of the way. It 
happened very often in Nevada that $90 to $100-ore 
was purchased for the purpose of amalgamating it 
in pans; but a few tons proved that amalgamation 
had to be given up. Such ore is now considered 
suitable only for smelting. 

At a very trifling expense all these difficulties can 
be avoided and the amalgamation turned into a per- 
fect success ; for instance, the amalgamation of the 
silver ores at Flint, Idaho, (32) turned out such 
base amalgam that further working proved to be 
ruinous. The introduction of the leaching pro- 
cess, however, resulted in a most favorable amal- 
gamation. It is only to be regretted that after 
working several hundred tons, the mine refused to 
provide the mill with ore, perhaps on account of 
not having been sufficiently opened. The leaching 
for the pan amalgamation is most important and 
at the same time cheap ; all the expense is reduced 
to that of obtaining hot water. This process is 
not only important for silver ores containing base 
metals, but also for gold ores which by their na- 
ture require roasting. This refers principally to 
auriferous copper ores, as the amalgamation of gold 
is very much obstructed by the presence of copper 
salts. 



138 EXTRACTION OF GOLD. 

It is a matter of surprise how so simple a remedy 
could have been overlooked while fighting with the 
obstructions, caused by rebellious ores, during the 
amalgamation. If there is soluble chloride of sil- 
ver in the roasted ore, and besides this, soluble 
chlorides of copper, lead, antimony and zinc, it is 
a matter of course that all will be decomposed and 
amalgamated. All take part in consuming and 
parting the quicksilver, and in destroying the pan, 
hindering at the same time the easy amalgamation 
of the silver and gold. Why, then, not put all 
these obstructive metals out of the way and give the 
silver a better chance to amalgamate ? The base 
metal chlorides are soluble in water, the chloride 
of silver is not. It is therefore a most simple ma- 
nipulation to dissolve those salts in water and to 
remove them from the ore before amalgamation, by 
the leaching process. As soon as this is done the 
ore is divested of its rebellious nature and it be- 
haves in pans like the best ore. 

The process of leaching is described 61 a. 



IV. EXTRACTION OF GOLD. 

78. The extraction of gold without the use of 
quicksilver is limited mostly to those ores in which 
the gold is not free in a metallic condition, but 
combined with sulphur or arsenic in the respective 
pyrites. 



EXTRACTION OF GOLD. 139 

There is only one body with which the gold must 
be combined before subjected to further treatment, 
and this body is chlorine. The chlorination can be 
effected either during roasting, 38, or after roast- 
ing, by contact with chlorine gas, 74, 79, or finally 
by contact with chlorinated water, 73. The usual 
way of chloridizing the gold is that by introducing 
the chlorine gas into the roasted ore, when cold. 
This mode is described fully in Kustel's work on 
" Concentration and Chlorination." No improve- 
ment of importance has been since introduced in 
this process. 

The Chlorination Process. (Plattner's. ) 



79. This process is based on the property of 
metallic gold of being changed into a soluble 
chloride of gold when in contact witji chlorine gas. 
The chloride of gold can be dissolved in water, 
separated from the ore by lixiviation, and then 
precipitated in a metallic condition by a solution 
of sulphate of iron. There are several establish- 
ments in California, principally in Grass Valley, * 
where auriferous pyrites are treated by chlorina- 
tion on a large scale. By way of chlorination, if 
properly executed, 90 to 95 per cent, of the fire 
assay can be extracted. 



* The first idea of trying this process on sulphurets in California 
came from Mr. Ch. Von Beseler, who experimented on it with Mr. Deet- 
ken in 1858. Since then Mr. Deetken has been engaged in the process, 
up to the present day, superintending chlorination works in Grass Val- 
ley. 



140 



EXTRACTION OF GOLD. 



80. In order to be sure of a result on a large 
scale, it is an easy matter to make an experiment 
with twenty or thirty pounds of sulphurets or ore 
in the following way : The named quantity must 
be roasted first, and it is the most difficult task, 
requiring either a small furnace or a great deal of 
patience, especially when small charges are treated 
on a large piece of sheet iron, having a charcoal 
fire beneath. In either case the sulphur must be 
driven out perfectly, so that when in a glowing 
condition, no smell of sulphurous acid can be ob- 
served. When finished a scruple is taken for an 
assay, and the roasted stuff moistened with water, 
after the weight of the whole has been noted. 

Fig. 17. 




A common water bucket is then prepared to re- 
ceive the moistened ore, which must not be too wet, 
but only moist enough to allow its being sifted. 
On the bottom of the bucket, a, Fig. 17, some clean 
rock or broken glass is placed about two inches 
deep, and covered with a piece of moistened can- 
vas. A short glass pipe, c, two-eighths of an inch 
in diameter, is inserted close above the bottom. 



EXTRACTION OF GOLD. 141 

The ore, d, is then introduced, filling up two- 
thirds or less of the space as loosely as possible, 
and covered with a wooden or iron cover and 
pasted all around with dough. The cover is pro- 
vided with a short glass tube, like c, to which an 
india rubber tube,-/", for carrying the gas out of the 
room is attached. Both glass tubes, c and*/", must 
be likewise secured with dough. 

The chlorine gas is generated in a glass vessel, 
A*. There are two corks in it, each having a 
glass tube, as represented in the drawing. The 
cork, I, is removed and the vessel charged with 
3 ounces of peroxide of manganese, 4 ounces of 
common salt, and 4J ounces of water all of which 
are well mixed. The cork is inserted again and 
well secured with dough. Another vessel, B, pro- 
vided with two necks, contains water as indicated 
by (7; the glass tube, h, dips about one-half inch 
into the water. The corks are made air tight like 
the others in A. The whole apparatus is now joined 
together by rubber pipe, n and o, fitting tightly to 
the glass tubes. Having all thus prepared, 7-J 
ounces of sulphuric acid are poured through the 
safety-tube, m, but only in small portions and at 
intervals. When the bubbling of the water at g, in 
the vessel B, is not lively enough, some more acid 
is introduced, and finally the temperature raised 
by an alcohol lamp. If all the joints have been 
luted carefully with dough, not the slightest incon- 

* All materials necessary for such an apparatus can be procured from 
John Taylor, on Washington street, San Francisco. 



142 EXTRACTION OF GOLD. 

venience will be met with. The chlorine gas from 
the generator, A, is forced through the water in JB, 
by this means washed from muriatic acid. Through 
the pipe, o, it enters the bucket and ascends slowly 
till it reaches the cover, escaping then through the 
rubber pipe, F, where it must be examined from 
time to*time by dipping a glass rod into ammonia 
and holding it to the end of the pipe, x, which 
leads out of the room. In contact with chlorine 
the ammonia evolves white fumes, and chlorine can 
be detected by these means wherever it may escape. 
The gas is allowed to pass through the bucket as 
long as chlorine is created. In this condition, by 
stopping up the pipe, x, if no more chlorine is 
evolved the apparatus may stand until the next 
day. The cover is then removed, the pipe, o, taken 
off, a clean glass or porcelain vessel, as indicated 
by z, placed below c, and warm water carefully 
poured over the ore till the bucket appears to be 
full. The solution which comes out at c, must be 
examined at times in a small tumbler with a few 
drops of a solution of sulphate of iron. If the 
clear solution remains unchanged, without becom- 
ing darker, the lixiviation is finished. 

To the solution in the vessel, z, a few drops of 
muriatic acid and then sulphate of iron, or green 
vitriol, (dissolved) is added and stirred with a glass 
rod. The whole is allowed to stand till all the 
gold is precipitated and the liquid is perfectly 
clear. This is drawn off by means of a syphon, 
for which the rubber pipe, x, can be used. The 



EXTRACTION OF GOLD. 143 

remaining fluid and the precipitated gold is gath- 
ered on a filter, washed with warm water and dried 
with the filter in a porcelain cup, above an alcohol 
lamp. The filter is burned either free or under a 
muffle, care being taken not to lose a particle of the 
filter ashes ; mixed with some lead it is then cu- 
pelled and the gold button weighed. A compari- 
son with the assay shows to what percentage the 
chlorination has proceeded. 

Chlorination of Sulphurets and Arseniu- 
rets. 

8 1 . All the proceedings of the chlorination have 
been treated already in describing the process of 
silver to which we will here refer. The first oper- 
ation is an oxidizing roasting, as explained in 44. 
When roasted, the ore is moistened, charged into 
chlorinating vats which are preferable to boxes for 
gold and chloridized according to 74. After sev- 
eral chlorinations have been performed and the gold 
has accumulated in the precipitating vat, the inside 
of which should be varnished with asphaltum var- 
nish, the gold is taken out by means of a scoop, 
put into a clean porcelain dish or enameled vessel, 
filtered, washed first with diluted nitric acid and 
then with hot water, dried and melted with the ad- 
dition of some borax. 

Other methods of extracting gold without mer- 
cury are mentioned in 38 and 73. 



TABLE OF CONTENTS. 



I. INTRODUCTION. 

PAGE. 

Classification of Ores 5 

Important Silver Ores 5 

Difference between Beal Silver Ores and Argentiferous 

Ores 7 

Important Combinations 8 

Means of Desulphurization 8 

Eesult of Desulphurization 11 

Means of Reduction , , 12 

Desulphurization of Ores not Efficient 13 

What a Chloride is, and how Chlorination is effected. ... 14 

Means of Separating the Metals from Chlorine 16 

II. BOASTING OF ORES. 

A. Chloridizing Boasting 22 

Necessary amounts of Sulphurets 26 

Amount of Salt used 28 

Permanent 'stirring not essential 31 

Signs of a good Chloridizing Boasting 32 

Means of destroying Base Metal Chlorides 35 

Steam decomposes Base Metal Chlorides 37 

Application of Steam in Boasting 37 

Lead has a bad influence 38 

Difference in Boasting Processes 39 

Examples of Boasting Processes 39 

In what condition the Metals are after Boasting 41 

Charges in Boasting 67 



TABLE OF CONTENTS. Cxlv 

PAGE. 

B. Oxidizing Roasting 67 

Chemical Changes in Roasting 68 

What Process Requires Oxidizing Roasting 69 

Roasting Furnaces 76 

Furnaces Managed by Handwork 80 

Reverberatory Furnaces 80 

Single Roasting Furnace 80 

Double Roasting Furnace " 84 

Long Roasting Furnace 85 

Muffle Furnace 89 

Furnaces with Mechanical Apparatus 90 

Revolving Hearth Furnace 91 

Ernst's Rotary Furnace 93 

Parke's Furnace 95 

Bruckner's Furnace 96 

O'Hara's Chain Furnace 97 

Stedefeldt's Furnace 99 

Chimneys and Flues 102 

III. EXTRACTION OF SILVER BY LIXIVIATION. 

Solving Process 104 

Extraction of Silver 105 

Precipitation of the Silver Ill 

Treatment of Precipitated Silver 113 

Precipitation of Copper 115 

Quality of Ores Fit for the Solving Process 121 

Sulphide of Calcium 123 

Hyposulphite of Lime 125 

Patera Process 126 

Kiss Process 126 

Patera and Roszner Process 127 

Kustel and Hofmann Process 128 

Augustin Process 132 

Ziervogel Process 134 

The Leaching Process 136 

IV. EXTRACTION OF GOLD. 

The Chlorination Process (Plattner's).. . 139 

Chlorination of Sulphurets a.jjj^zf^iRi^pps^^ 143 

OF THB 

UNIVERSITY 



ETURN CIRCULATION DEPARTMENT 

O * 202 Main Library 


DAN PERIOD 1 
HOME USE 


2 


3 




5 


6 



ALL BOOKS MAY BE RECALLED AFTER 7 DAYS 

Renewals and Recharges may be made 4 days prior to the due date. 

Books may be Renewed by calling 642-3405. 



DUE AS STAMPED BELOW 


INTERUBRARY LOAN 




APR fi, 198 




miv AC r AI NT 

VwWT* %r" %^^- ' * - , LZ ' 


pg^ 





















































ORM NO. DD6 



UNIVERSITY OF CALIFORNIA, BERKELEY 
BERKELEY, CA 94720 



YB 38355 






UNIVERSITY OF CALIFORNIA LIBRARY